CA2072629C - User-interactive reduction of scene balance failures - Google Patents
User-interactive reduction of scene balance failuresInfo
- Publication number
- CA2072629C CA2072629C CA002072629A CA2072629A CA2072629C CA 2072629 C CA2072629 C CA 2072629C CA 002072629 A CA002072629 A CA 002072629A CA 2072629 A CA2072629 A CA 2072629A CA 2072629 C CA2072629 C CA 2072629C
- Authority
- CA
- Canada
- Prior art keywords
- image
- adjustment mechanism
- digitized
- image adjustment
- discriminator
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/46—Colour picture communication systems
- H04N1/56—Processing of colour picture signals
- H04N1/60—Colour correction or control
- H04N1/6011—Colour correction or control with simulation on a subsidiary picture reproducer
-
- G—PHYSICS
- G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
- G03B—APPARATUS OR ARRANGEMENTS FOR TAKING PHOTOGRAPHS OR FOR PROJECTING OR VIEWING THEM; APPARATUS OR ARRANGEMENTS EMPLOYING ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ACCESSORIES THEREFOR
- G03B27/00—Photographic printing apparatus
- G03B27/72—Controlling or varying light intensity, spectral composition, or exposure time in photographic printing apparatus
- G03B27/73—Controlling exposure by variation of spectral composition, e.g. multicolor printers
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
- Control Of Exposure In Printing And Copying (AREA)
Abstract
A photofinishing image processing operator detects failures in the performance of a scene balance mechanism on a digitized image and enables a photofinisher to interactively correct for such failures before the failed image has been processed and an unacceptable output image printed. Whenever a failed image is identified, one or more reasonably low resolution version of the image as processed by the scene balance mechanism are displayed to the photofinishing operator, together with a request for image adjustment information that is used to modify or correct the color balance of the image. In response to this user input information, the image processor adjusts color balance parameters of the scene balance mechanism, so that the digitized image processed by the modified scene balance mechanism will yield a print of acceptable color balance quality. This interaction of the photofinishing operator with the image processing system may be accomplished by way of several modes, including displaying differently processed versions of the image, or a single version of the image processed by the scene balance mechanism, together with a question regarding the type of image. The failure detection mechanism preferably employs a non-linear discriminator, such as a neural network.
Description
WO ~2/0~47n PCT/lJS91/06577 . 1 2~72~29 USER- INTERACTIVE R~;LJU~; l lON OF SCENE
RP.T.A~'~ p'ATT.~P~.S~
.
FI~T~ OF T~ INV~TIQN
The present invention relates in general to digitized color imagery photofinishing systems and is particularly directed to a mechani.sm for detecting failures in the performance of an im~ge adjustment (e.g. scene balance) mechanism and, in response to a detected potential failure, interactively adjusting color balance control parameters o~ the image adjustment mechanism, so as to enable that digitized color image to be modified or corrected (e.g. color balanced) in a manner that will produce a representation that is pleasing to the human visual system.
~CKGROUND OF T~ INV~TION
Optical photofi n i sh;ng systems customarily employ some form of image correction or adjustment mechanism, such as a scene balance mechanism, which - evaluates the content of each digitized image and selectively adjusts or modifies the image, for example adjusts its color balance, so that, when reproduced, as by way of an optical printer, the printed image will be a high quality color reproduction that is pleasing to the human visual system. For purposes of the present invention, ~y 'scene balance mechanism~ is meant an adjustment of image color bal~nce ~ased upon the scene content and the sensitometric characteristics (e.g.
exposure, light source, etc.) of the im~ge being reproduced. Adjustment of ~color balance' is defined as adjusting the average red, green and blue image levels, so as to ensure that an image will ha~e the appropriate 3~ color and neutral reproduction characteristics.
, . . .. . . .... . . . . . .. . . . .. . . . . . . .
...... . . . . . . . . . . . . . .. . . . . .
~ . .. . . ... .
.. . .
', . ' . . .
Wos2/0s47n PCT/VS91/06~77 2 ~ 71~J~,2 ~
Unfortunately, because su~h image adjustment mechanisms are typically statistically based, some images will be unsuccessfully treated and, consequently, will yield an unacceptable reproduction.
Moreover, depending upon the particular image correction mechanism employed, the rate at whi~h images are unsuccessfully processed can be significant (e.g.
on the order of five percent). In those instances where the correction mechanism fails to properly correct the image ~e.g. adjust its color balance), either the customer receives a poor guality print or the photofinisher discovers the mistake and then manually corrects the image. In either case (in response to a disgruntled customer or a predelivery intercept during a photofinisher's review), because the error is not corrected until after the image is printed, the photofinisher incurs a substantial cost in materials (the print paper being a major portion of the cost of film processing) and time (due to the break in the flow of the image reproduction process).
SUM~RY OF TH~ INV~TIQN
In accordance with the present invention, rather than simply tolerate a failure in the performance of the image correction mechanism and manually correct the improperly processed image after an unacceptable print has been discovered, an image adjustment failure detection mechanism is inserted into the photofinishing process upstream of the reproduction (printing) stage, in order to identify those images where the image correction mechanism is likely to fail.
Once such a 'likely to fail1 image is identified, one or more r asonably low resolution versions of the digitized image, produced as a result of modifications . "
wos2/05470 PCT/US91/06577 7 2 ~ 2 ~
or parameter corrections of the adjustment mechanism are displayed to the photofin}sher, together with a request for the photofinisher to input image adjustment information that is used to modify or correct (the color balance of) ~he image. In response to this user input information, the imaye processor adjusts parameters of the image correction mechanism, so that the digitized image processed by the selected (modi~ied) image adjustment mechanism will yield a reproduction (print) of acceptable color balance quality. This int~raction of the photofinishing operator with the image processing system may entail several modes. Where the image adjustment mechanism is a scene balance mechanism for balancing the color content of the image, each mode serves to modify control parameters o~ the scene balance operation.
In accordance with a first mode, plural, low resolution versions of the processed image of acceptable presentation quality are displayed on the photofinisher~s workstation display. One of these image versions (e.g. one that has been centrally placed within an array of images across the display) shows how the digitized image would appear based upon the current response of the image adjustment (scene balance) mechanism. Displayed versions of the image on opposite sides of this one im~ge show ~ariations in color balancP resulting from iteratively shifted parameter adjustments to the (scene balance) mechanism (e.g.
lighter and darker where the failure mode detected is one of neutral b~lance). By way of a display prompt, the photofinisher is r~ ested to choose from among the different displayed vel ions of the im~ige the one that is most pleasing. In response to this selection, the manner in which the scene balance mechanism processes - .
.. . .... . . ......... . . .
. ~ .: . ~ . - - -- , - ~
WOg2/05470 P~T/US91/06577 ~ 4-the digitized image is modified to conform with the color balance characteristics of the selected image version.
Pursuant to a second mode of operation, which is similar to the first mode, the user is given a choice of only two versions of the im~ge - the one currently produced by t~he scene balance mechanism and an alternative that is based upon a set of modified parameters. As in the first mode, the digitized image is then processed to conform with the color balance characteristics of the selected image version, indicating one of two possible scene interpretations, such as flash vs. non-flash or color- subject failure lS vs. non-subject failure.
In accordance with a third mode of operation, a low resolution version of the image processed by the scene balance mechanism is displayed to the photofinishing operator, together with a question regarding the type of im ge, e~g. whether or not the image is a backlit scene or a flash scene, which may seem similar to the scene balance mechanism but require different balance processing. The photofinisher's response is then employed to refine how the scene balance mechanism corrects the ima~e (e.g. corrects for an illumination ambiguity).
The mechanism for detecting failures in the performance of the scene balance mechanism may comprise a statistically based, linear discriminator, such as a Fischer Discr;minAnt or a Bayes Classifier, which employ a multi~,m~n~ional discrimination space derived from predictors computed from the imagery data.
Preferably, however, the fallure detection mechanism .
. :
- -W092/0~470 PCT/US91/06577 ~72~2~
employs a non- linear discriminator, such as a back propagation-trained n~ural network, because of its - ability to accurately recognize images that the scene balance mechanism will correctly process. Typically, both linear discriminators and neural networks will detect potential color balance failures for about 50~
of poor ~uality im~ges; however, the preferred neural .
network-based failure discriminator will effectively never detect a potential failure for a correctly color-~0 balanced image~ .
BRI~F D~SCRIPTION OF T~ DRAWINGS
Figure l dia~ atically illustrates aphotographic color film photofinishing minilab with ~5 which the image adjustment (e.g. scene balance) failure detection and correction mechanism of the prPsent invention may be employed;
Fiaure 2 diay~lullatically shows an example of a neural network-configured scene balance failure discriminator containing a feed-forward n-layer perceptron (trained by back propagation);
Figure 3 illustrates a first 'sliding scale' mode of the failure detection and interactive correction process according to the present invention;
Figure 4 illustrates a second mode ~f the failure detection and interactive correction process according to the present invention in which a photofinishing minilab operator is given a choice of two differently scene balance processed versions of an image;
. - -.. . . . .
.. .... . . .
... ...
.. , , . , . _ . . :i :-. -:; . .
: . :: , ~, : . . .
W092/05470 P~T/~S91/06577 ~o~ 2$~ ~ -6-Figure 5 diagrammatically illustrates a third mode of operation of the scene balance failure discriminator, wherein a low resolution version of the image processed by the scene balance mechanism is displayed to the photofinishlng operator, together with a question regarding the type of ima~e; and Figure 6 diagrammatically illustrates a multiregion map used to define the characteristics of a scene balance failure discriminator.
~TAIT.~ DESCRIPTION
Before describing in detail the image adjustment failure discrimination mechanism in accordance with the present invention, it should be obsèrved that the present invention resides primarily in a novel structural combination of conventional digital imagery processing circuits and com2onents and not in the particular detailed configurations thereof Accordingly, the structure, control and arrangement of these conventional circuits and components have been illustrated in the drawings by readily understandable block diagrams which show only those specific details that are pertinent to the present invention, 50 as not to obscure the disclosure with structural details which will be readily apparent to those skilled in the art having the benefit of the description herein. Thus, the block diagram illustrations of the Figures do not necessarily represent the mechanical structural arrangement of the exemplary system, but are primarily intended to illustrate the major structural components of the system in a convenient functional grouping, whereby the present invention may be more readily understood.
...... . . . .. . . . . . .. . . . . . .
~ -; . , . :. .. . . .
. :- , ., : . .. :;-. -. .
- , . ~.: :,.. ~,, -. . . .
~: .......... .... . . . ..
W092/0~470 PCT/US91/06577 _ -7-~72~29 Figure 1 diagramm~tically illustrates a photographic color film processing system (e.g.
photofinishing minilab) with which the im2ge adjustment failure discriminator in accordance with the present invention may be employed and, for purposes of the present description, such a system m~y be of the type described in co-pending Patent application Serial Number , filed , by S. Kristy, entitled aMulti-resolution Digital Imagery Photofirlishing System~, assigned to the assignee of the present application and the disclosure of which is incorporated herein. It should be observed, however, that the system described in the above-referenced co-pending Kristy application is merely an example of one type of system in which the invention may be used and is not to be considered limitative of the invention. In general, the invention m~y be incorporated in any digitized imagery processing and reproduction system. Moreover, although, in the description to follow, the failure discriminator will be describ~d in terms of the opera~ion of a scene balance mech~nism, it should be observed that the principles presented here are applica~le to other image adjustment schemes.
~5 In accordance with the digital image processing system of Figure 1, photographic color images, such as a set of twenty-four or thirty-six 36mm-by-24mm im~ge frames of a 35m~ film strip 10, ar~
scanned by a high resolution opto-electronic color film scanner 12, such as a commercially availa~ie Eikonix Model 1435 scanner. High resolution film scanner 12 outputs digitally encoded data representative of che response of its imaging sensor pixel array ~e.y. a 3072 x 2048 pixel matrix) onto which a respective photographic image frame of film strip 10 has been - :. - - , . .
W~92/~70 PCT/US91/06577 2~
projected by an input imaging lens ~ystem. This digitally encoded data, or 'digitized' image, is encoded to some prescribed resolution (e.g. 16 bits per color per pixel) that encompasses a range of values over which the contents of the scene on the color film may va~y. For a typical color photographic negative, the range of valu~s i~ less than the density vs.
exposure latitude of the film, but is sufficiently wide to encompass tho~e density values that can be expected to be encountered for a particular scene. Then, by means of a scene balancing mechanism resident in an attendant image processing workstation 14, which contains a frame store and image processing application software through which the digitized image m~y be processed to achieve a desired base image appearznce and configuration, the digitized image is mapped into a set of lower bit resolution digital codes (e.g. 8 bits per color per pixel), each of which has a resolution corresponding to the dynamic range of a digitized image data base (e.g. frame store), the contents of which may be adjusted in the course of driving a reproduction device 16, for example, enabling a print engine to output a high quality color print.
Preferably, the digitized imagery data output by the high resolution film scanner is subjected to a code conversion mechanism of the t~e described in co-pending application Serial No. , filed , by Y.
Madden et al, entitled ~Extending Dynamic Range of Stored Image Database,~ assigned to the assignee of the present application and the disclosure of which is herein incorporated. Pursuant to this code conversion scheme, the dynamic range of the digitized image database is extended so as to permit shifting of encoded pixel values without 'clipping', and to provide . . .
. ~ .
.
:
. ~ : ' :
WOg~/0~7~ PCT/US91/06577 2~72~29 a limited window of values into which extremely high reflec~ance image points may be éncoded and stored.
Digital codes into which the image scanner output are mapped by the scene balance mechanism are converted into a set of reduced-range digital codes of the same resolution as, but having a smaller range o~ image content values than the dynamic range of the digitized image data base. The code conversion mechanism operates to convert a maximum value of 100% white reflectance to an encoded value that is less than the upper limit of the dynamic range of the database to accommodate shifts in the digitized imagery data and allow for the placement of specular highlights that are beyond the 100% white reflectance maximum.
In accordance with the imagery data processing system described in the above referenced co-pending Kristy application, each high resolution captured image is stored ~s a respective image data file cont~in;n~ a low, or base, resolution image bit m~p file and a plurality of higher resolution residual images associated with respecti~ely increasin~ degrees of image resolution. By iteratively combining these higher resolution residual images with the base resolution image, successively increased resolution images may be recovered from the base resolution image.
As an example, spatial data values representativP of a high resolution (3072 x 2D48) image scan of a 36mm-by- 24mm image frame of a 35mm film strip 10 may be stored as a respective im~ge data file including a base resolution im~ge bit map file containing data values associated with a spatial image array or matrix of 512 rows and 768 columns of pixels and an associated set of residual image files to be , ,. . , ,:
., .. . . . . .. . . .....
-- ~, .
.
-. .. - . - : : ~
.. - :, . :, ; ,:~ : :
W092/0~470 PCT/US9l/06577 ?,~ o-stored on the disc. Within the workstation itself, the base rPsolution image m~y be further sub- sampled to derive an even l~wer resolution sub-array of image values (e.g. o~ the order of 128 x 1~2 pixels) for use by the photofinishing operator in the course of formatting and storin~ a digitized image file.
Pursuant to the present inv~ntion, a substantially reduced resolution version of each di~iti2ed image to be color balance-processed by the scene balance mechanism is also coupled to a scene balance failure discriminator, which determines the likelihood that the output image processed by the scene balance mechanism will be of acceptable color balance quality. For this purpose, the scene balance mechanism failure discriminator P~m;neS the spatial content of a very small sub-array (e.g. 24 x 36 pixels) produced by subsampling compression of a 128 x 192 pixel version of the image of interest and, based upon a previously generated statistical model of prescribed characteristics of the image, derives a measure of the likelihood that the color balance performance of the scene balance mechanism for th~t image will produce a visuall~ acceptable reproduction (high resolution output print).
The discriminator, per se, may comprise a statistically based, linear discriminator, such as a Fischer Discrimin~nt or a Bayes Classifier, which use a multi~;men~ional discrimination space derived from predictors computed from the imagery data. Preferably, however, the failure detection mechanism employs a non-linear discriminator, such as a back propagation-trained neural network, because of its ability to ~5 accurately recognize images that will be correctly .
.. . . .. .
.. . , . - . - - . - --- . ~ :
.. . . ..
.. ... .
, . .
-, , . :. . .,- .
.
2~372~29 processed by the scene balance mechanism. Namely, although both linear discriminators and neural networks are capable of detecting color balance failures for about 50% of the images that will fail ('bad' im~ges), a linear discriminator will also erroneously flag failùres for some images that will be correctly processed by the scene balance mechanism. On the other hand, the preferred neural network-based failure discriminator will effectively never detect a scene balance failure fc)r a 'good' image.
The substantially reduced resolution version of the imaye, such as one comprising â pixel sub-array in the order of only 24x36 pixels, contains sufficient ~5 spatial color content to permit identification of scene balance failures for a high resolution image, but with considerably reduced computational intensity. Still, even with such a reduced image size, the number of values (864 for a 24 x 36 pixel sub-array) that enter into the statistical analysis is still substantial and, for a linear discriminator, may have an associated set of approximately 50 statistical measures (predictors) of the content of the scene. For the same sized image sub- array, it is possible to build a back propagation neural network that operates on a reasonably constrained set of parameters.
As one non-limitative example, a neural network, such as that diay~.u,~tically shown in Figure 2 as a feed-forward n-layer perceptron (trained by back propagation), may be configured of a nine element input layer 31, a twel~e element hidden layer and a single element output layer 33. An image training population on the order of several thousand 24x36 pixel scenes repeatedly and successively applied to the network m~y ~ , . . . .
... .
.: .. . ... - ~ j: . ,... :
-- :: .. . ~ .. .
. ,, ~ , . , . , , .. :. ~:
wos~/0~7~ q~ -12- P~T/US91/06;77 be used to load the network. As a non-limitative example, a set of nine predictors (Nos. l-9) listed in Table l, below and referencing the region map of Figure 6, m~y be employed. It should be observed, however, that other predictors in addition to and different from those listed here, m~y be used. For a discussion of the use of neural networks for pattern recognition and choosing predictors, attention may be directed to the text: ~Adaptive Pattern Recognition and Neural Netwoxks~, by Yoh-Haw Pao, Chapters l, 2 and 5.
T~RT,F~ 1 p~nI~ToR D~FINI~ION
No. l Red maximum of region 5 - red average of entire frame.
No. 2 Red maximums of regions 3, 4, 5 are averaged together and the red average of the entire frame is subtracted from it.
No. 3 Variance of neutral T space value for entire range, where T space is defined, for example, as set forth in U S.
Patents Nos. 4,159,174 and 4,g33,709).
No. 4 Neutral T space average of entire frame.
No. 5 Hyperbolic tangent of center maximumf minus the center ave.age.
No. 6 Neutral T space average of region 5 minus neutral T space average of regions l, 2, 3, 4.
N0. 7 Average of neutral T space maximums of regions 3, 4, 5 minus neutral T space average of entire frame.
No. 8 Neutral T space centroid of entire frame.
No. 9 Green magenta average of entire frame.
.. .. - , - - - -. ,:
W0~2/0~7~ PCT/US91/06577 -13- 2~72~2~
Moreover, the type of neural network employed - is not constrained to a feed-forward, back propagation implementation. Other network types, such as a counter propagation network may be used. Suffice it to say that examples given here have been shown to correctly identify scene balance failures for approximately 50%
of 'bad' images (ima~es for which the scene balance mechanism will fail) but 0~ of 'good' images, namely those images for which the scene balance mechanism will produce a properly color-balanced output image. For a further discussion of neural networks in general and their application to statistica} processing, attention m~y be directed to the text: ~Parallel Distributed Processing~ Vol. I-II, by McClelland, Rumelhart and PDP
Research Group, MIT Press, 1988.
The scene balance failure detection mechanism of the present invention is preferably incorporated within the image processing software employed by workstation 14, in order that those images where the scene balance mechanism is likely to fail will be identified and corrected prior to the reproduction (printing) stage of the photofinishing process. Once such a scene balance failure image is identified, one or more reasonably low resolution versions of the digitized image processed by the scene balance mech~nism are displayed to the photofinishing operator, tog~ ner with a request for the photofinishing operator 3Q to input image adjustment information that is used to modify or correct the color balance action of the scene balance mechanism on that im~ge. In response to this user input information, ~he image processor adjusts color balance parameters of the scene balance mechanism, so that the resulting processed image will . . .
: , , - -: - . , : , .", ., . :: .: : :. . ; .
: . .. , . -, ~ , :. :
WVg/0;47~ PC~/US~110677 produce a print of acceptable color balance quality.
This interaction of the photofinishing operator with the image processing system may be accomplished using any one of several modes dia~.rammatically illus~rated in Figure 3-5, described;below.
More particulaxly, Figure 3 illustrates a first mode of the failure detection and interactive correction process according to the present invention in which a plurality of low resolution (128 x 192 pixels) images 20-l, 20-2, 20-3,..., 20-N, each of which is a respectively different version of the (l28 x 192) digitized image a~ processed by the resident scene balance mechanism, are displayed as a linear array on the photofinisher's workstation display. The image tha~
results from the application of the currently defined scene balance mechanism to the scanned image may correspond to image 20-3, centrally placed within a linear array of images across the workstation display screen. Images 20-l, 20-2 and im~ges 20-4 and 20-5 on opposite sides of this image have respectively different variations in color balance obtained by successively offsetting or shifting the scene balance mechanism along a prescribed axis (e.g. lighter and darker scan densities where the failure mode detected is one of neutral balance~. Further variations in the image may be formed by the use of a conventional graphics processing scrolling operator to effectively create a 'sliding scale' for a linear array of images that may be panned across the screen.
By way of a display prompt, the photofinishing operator is requested to choose the most visually pleasing one of the different versions of the image within the display window associated with .
- . -: .
- - : . . , ~~ -15- ~ ~ 2 ~ 2~
respecti~ely different biasing of the sc-ne balance mechanism. In response to this selection, the manner in which the scene balance mechanism processes the digiti~ed image is modified to conform with the color - S balance characteristics of the selected image version.
For this purpose, the bias settings employed to produce the selected image are used to modify the scene balance mechanism. Namely, since each image version represents a prescribed offse~ or shift of th~ scene balance mechanism, selecting a given image effectively ' specifies the magnitude of the offset bias to be used by the scene balance mechanism in processing the high resolution image for storage and subsequent reproduction.
~5 Figure 4 illustrates a second mode of the failure detection and interactive correction process according to the present invention in which the photofinishing minilab operator is given a choice of only two versions of the image. As shown in the figure, on the workstation display two low resolution images 41 and 42, each of which is a respectively dif erent version of the (128 x 192) digitized image as processed by the resident scene balance mechanism, are displayed on the photofinisher's workstation display. One of the displayed images has been processed as though the image is front-lit (e.g. by a flash illumination source), while the other displayed ima~e has been processed as though the image is back-lit ~silhouette). Both types of images look similar to the scene balance mechanism, but -equire different processing. The photofinisher~s choice of which displayed version of the image is most pleasing determines which version of the scene balance mechanism will be used to process that image.
.. . . . . . ...
. ~ :
:, , -w~92/0~70 PCT/US91/06~77 Figure 5 diagrammaticallY illustrates a third mode of operation of the scene balance failure discriminator, whereln a low resolution version 51 of the image processed by the scene balance mechanism is displayed to the photofinishing operator, together with a panel 52 containing a question regarding the type of image, e.g. whether cr not the image is backlit or a flash scene. The photofinisher's input response to the displayed que~tion panel 52 ~e.g. by mouse or keyboard entry) is then employed by the image processing software to refine how the scene balance mechanism corrects the image (e.g. corrects ~or an ambiguity in the type of scene illumination). For example, for the case where the photofinisher enters the fact that the subject/scene is backlit, a scene balance mechanism optimized for a silhouette is employed. Similarly, for the case where the photofinisher enters the fact that the subject/scene is illuminated by a flash source, a scene balance mechanism optimized for a front-lit image is used.
As will be appreciated from the foregoing description, the present invention provides a technique for detecting images prone to scene balance failures and enabling the photofinisher to interactively correct for failures in the perform~nce of ~he scPne balance mechanism before an image has been processed and an unacceptable output im2ge printed. By correcting for a discovered scene balance failure prior to printing the image, the photofinisher will not incur what can amount to substantial penalties in terms of cost of materials and film processing time.
Moreover, the present invention offers the advantage that, by using a failure discriminator, only .
,- :
.:
: . ,-: ' --- :, . .
. ..
,, :.... : --wn g~/05470 PCT/US9l/~6577 2~262~
some, but not all, of the scenes need to by evaluated by the photofinisher. Without it, to detect as many failures as possible, the photofinisher would have to look at every im~ge.
While we have shown and described an embodiment in accordance with the present invention, it is to be understood that the same is not limited thereto but is susceptible to numerous changes and modifications as known to a person skilled in the art, and we therefore do not wish to be limited to the details shown and described herein but intend to cover all such changes and modifications as are obvious to one of ordinary skill in the art.
.. ..
_ . . . .. . . . . ... . . . . . . . . .. ..
, :: .: ' . ' :, . r
RP.T.A~'~ p'ATT.~P~.S~
.
FI~T~ OF T~ INV~TIQN
The present invention relates in general to digitized color imagery photofinishing systems and is particularly directed to a mechani.sm for detecting failures in the performance of an im~ge adjustment (e.g. scene balance) mechanism and, in response to a detected potential failure, interactively adjusting color balance control parameters o~ the image adjustment mechanism, so as to enable that digitized color image to be modified or corrected (e.g. color balanced) in a manner that will produce a representation that is pleasing to the human visual system.
~CKGROUND OF T~ INV~TION
Optical photofi n i sh;ng systems customarily employ some form of image correction or adjustment mechanism, such as a scene balance mechanism, which - evaluates the content of each digitized image and selectively adjusts or modifies the image, for example adjusts its color balance, so that, when reproduced, as by way of an optical printer, the printed image will be a high quality color reproduction that is pleasing to the human visual system. For purposes of the present invention, ~y 'scene balance mechanism~ is meant an adjustment of image color bal~nce ~ased upon the scene content and the sensitometric characteristics (e.g.
exposure, light source, etc.) of the im~ge being reproduced. Adjustment of ~color balance' is defined as adjusting the average red, green and blue image levels, so as to ensure that an image will ha~e the appropriate 3~ color and neutral reproduction characteristics.
, . . .. . . .... . . . . . .. . . . .. . . . . . . .
...... . . . . . . . . . . . . . .. . . . . .
~ . .. . . ... .
.. . .
', . ' . . .
Wos2/0s47n PCT/VS91/06~77 2 ~ 71~J~,2 ~
Unfortunately, because su~h image adjustment mechanisms are typically statistically based, some images will be unsuccessfully treated and, consequently, will yield an unacceptable reproduction.
Moreover, depending upon the particular image correction mechanism employed, the rate at whi~h images are unsuccessfully processed can be significant (e.g.
on the order of five percent). In those instances where the correction mechanism fails to properly correct the image ~e.g. adjust its color balance), either the customer receives a poor guality print or the photofinisher discovers the mistake and then manually corrects the image. In either case (in response to a disgruntled customer or a predelivery intercept during a photofinisher's review), because the error is not corrected until after the image is printed, the photofinisher incurs a substantial cost in materials (the print paper being a major portion of the cost of film processing) and time (due to the break in the flow of the image reproduction process).
SUM~RY OF TH~ INV~TIQN
In accordance with the present invention, rather than simply tolerate a failure in the performance of the image correction mechanism and manually correct the improperly processed image after an unacceptable print has been discovered, an image adjustment failure detection mechanism is inserted into the photofinishing process upstream of the reproduction (printing) stage, in order to identify those images where the image correction mechanism is likely to fail.
Once such a 'likely to fail1 image is identified, one or more r asonably low resolution versions of the digitized image, produced as a result of modifications . "
wos2/05470 PCT/US91/06577 7 2 ~ 2 ~
or parameter corrections of the adjustment mechanism are displayed to the photofin}sher, together with a request for the photofinisher to input image adjustment information that is used to modify or correct (the color balance of) ~he image. In response to this user input information, the imaye processor adjusts parameters of the image correction mechanism, so that the digitized image processed by the selected (modi~ied) image adjustment mechanism will yield a reproduction (print) of acceptable color balance quality. This int~raction of the photofinishing operator with the image processing system may entail several modes. Where the image adjustment mechanism is a scene balance mechanism for balancing the color content of the image, each mode serves to modify control parameters o~ the scene balance operation.
In accordance with a first mode, plural, low resolution versions of the processed image of acceptable presentation quality are displayed on the photofinisher~s workstation display. One of these image versions (e.g. one that has been centrally placed within an array of images across the display) shows how the digitized image would appear based upon the current response of the image adjustment (scene balance) mechanism. Displayed versions of the image on opposite sides of this one im~ge show ~ariations in color balancP resulting from iteratively shifted parameter adjustments to the (scene balance) mechanism (e.g.
lighter and darker where the failure mode detected is one of neutral b~lance). By way of a display prompt, the photofinisher is r~ ested to choose from among the different displayed vel ions of the im~ige the one that is most pleasing. In response to this selection, the manner in which the scene balance mechanism processes - .
.. . .... . . ......... . . .
. ~ .: . ~ . - - -- , - ~
WOg2/05470 P~T/US91/06577 ~ 4-the digitized image is modified to conform with the color balance characteristics of the selected image version.
Pursuant to a second mode of operation, which is similar to the first mode, the user is given a choice of only two versions of the im~ge - the one currently produced by t~he scene balance mechanism and an alternative that is based upon a set of modified parameters. As in the first mode, the digitized image is then processed to conform with the color balance characteristics of the selected image version, indicating one of two possible scene interpretations, such as flash vs. non-flash or color- subject failure lS vs. non-subject failure.
In accordance with a third mode of operation, a low resolution version of the image processed by the scene balance mechanism is displayed to the photofinishing operator, together with a question regarding the type of im ge, e~g. whether or not the image is a backlit scene or a flash scene, which may seem similar to the scene balance mechanism but require different balance processing. The photofinisher's response is then employed to refine how the scene balance mechanism corrects the ima~e (e.g. corrects for an illumination ambiguity).
The mechanism for detecting failures in the performance of the scene balance mechanism may comprise a statistically based, linear discriminator, such as a Fischer Discr;minAnt or a Bayes Classifier, which employ a multi~,m~n~ional discrimination space derived from predictors computed from the imagery data.
Preferably, however, the fallure detection mechanism .
. :
- -W092/0~470 PCT/US91/06577 ~72~2~
employs a non- linear discriminator, such as a back propagation-trained n~ural network, because of its - ability to accurately recognize images that the scene balance mechanism will correctly process. Typically, both linear discriminators and neural networks will detect potential color balance failures for about 50~
of poor ~uality im~ges; however, the preferred neural .
network-based failure discriminator will effectively never detect a potential failure for a correctly color-~0 balanced image~ .
BRI~F D~SCRIPTION OF T~ DRAWINGS
Figure l dia~ atically illustrates aphotographic color film photofinishing minilab with ~5 which the image adjustment (e.g. scene balance) failure detection and correction mechanism of the prPsent invention may be employed;
Fiaure 2 diay~lullatically shows an example of a neural network-configured scene balance failure discriminator containing a feed-forward n-layer perceptron (trained by back propagation);
Figure 3 illustrates a first 'sliding scale' mode of the failure detection and interactive correction process according to the present invention;
Figure 4 illustrates a second mode ~f the failure detection and interactive correction process according to the present invention in which a photofinishing minilab operator is given a choice of two differently scene balance processed versions of an image;
. - -.. . . . .
.. .... . . .
... ...
.. , , . , . _ . . :i :-. -:; . .
: . :: , ~, : . . .
W092/05470 P~T/~S91/06577 ~o~ 2$~ ~ -6-Figure 5 diagrammatically illustrates a third mode of operation of the scene balance failure discriminator, wherein a low resolution version of the image processed by the scene balance mechanism is displayed to the photofinishlng operator, together with a question regarding the type of ima~e; and Figure 6 diagrammatically illustrates a multiregion map used to define the characteristics of a scene balance failure discriminator.
~TAIT.~ DESCRIPTION
Before describing in detail the image adjustment failure discrimination mechanism in accordance with the present invention, it should be obsèrved that the present invention resides primarily in a novel structural combination of conventional digital imagery processing circuits and com2onents and not in the particular detailed configurations thereof Accordingly, the structure, control and arrangement of these conventional circuits and components have been illustrated in the drawings by readily understandable block diagrams which show only those specific details that are pertinent to the present invention, 50 as not to obscure the disclosure with structural details which will be readily apparent to those skilled in the art having the benefit of the description herein. Thus, the block diagram illustrations of the Figures do not necessarily represent the mechanical structural arrangement of the exemplary system, but are primarily intended to illustrate the major structural components of the system in a convenient functional grouping, whereby the present invention may be more readily understood.
...... . . . .. . . . . . .. . . . . . .
~ -; . , . :. .. . . .
. :- , ., : . .. :;-. -. .
- , . ~.: :,.. ~,, -. . . .
~: .......... .... . . . ..
W092/0~470 PCT/US91/06577 _ -7-~72~29 Figure 1 diagramm~tically illustrates a photographic color film processing system (e.g.
photofinishing minilab) with which the im2ge adjustment failure discriminator in accordance with the present invention may be employed and, for purposes of the present description, such a system m~y be of the type described in co-pending Patent application Serial Number , filed , by S. Kristy, entitled aMulti-resolution Digital Imagery Photofirlishing System~, assigned to the assignee of the present application and the disclosure of which is incorporated herein. It should be observed, however, that the system described in the above-referenced co-pending Kristy application is merely an example of one type of system in which the invention may be used and is not to be considered limitative of the invention. In general, the invention m~y be incorporated in any digitized imagery processing and reproduction system. Moreover, although, in the description to follow, the failure discriminator will be describ~d in terms of the opera~ion of a scene balance mech~nism, it should be observed that the principles presented here are applica~le to other image adjustment schemes.
~5 In accordance with the digital image processing system of Figure 1, photographic color images, such as a set of twenty-four or thirty-six 36mm-by-24mm im~ge frames of a 35m~ film strip 10, ar~
scanned by a high resolution opto-electronic color film scanner 12, such as a commercially availa~ie Eikonix Model 1435 scanner. High resolution film scanner 12 outputs digitally encoded data representative of che response of its imaging sensor pixel array ~e.y. a 3072 x 2048 pixel matrix) onto which a respective photographic image frame of film strip 10 has been - :. - - , . .
W~92/~70 PCT/US91/06577 2~
projected by an input imaging lens ~ystem. This digitally encoded data, or 'digitized' image, is encoded to some prescribed resolution (e.g. 16 bits per color per pixel) that encompasses a range of values over which the contents of the scene on the color film may va~y. For a typical color photographic negative, the range of valu~s i~ less than the density vs.
exposure latitude of the film, but is sufficiently wide to encompass tho~e density values that can be expected to be encountered for a particular scene. Then, by means of a scene balancing mechanism resident in an attendant image processing workstation 14, which contains a frame store and image processing application software through which the digitized image m~y be processed to achieve a desired base image appearznce and configuration, the digitized image is mapped into a set of lower bit resolution digital codes (e.g. 8 bits per color per pixel), each of which has a resolution corresponding to the dynamic range of a digitized image data base (e.g. frame store), the contents of which may be adjusted in the course of driving a reproduction device 16, for example, enabling a print engine to output a high quality color print.
Preferably, the digitized imagery data output by the high resolution film scanner is subjected to a code conversion mechanism of the t~e described in co-pending application Serial No. , filed , by Y.
Madden et al, entitled ~Extending Dynamic Range of Stored Image Database,~ assigned to the assignee of the present application and the disclosure of which is herein incorporated. Pursuant to this code conversion scheme, the dynamic range of the digitized image database is extended so as to permit shifting of encoded pixel values without 'clipping', and to provide . . .
. ~ .
.
:
. ~ : ' :
WOg~/0~7~ PCT/US91/06577 2~72~29 a limited window of values into which extremely high reflec~ance image points may be éncoded and stored.
Digital codes into which the image scanner output are mapped by the scene balance mechanism are converted into a set of reduced-range digital codes of the same resolution as, but having a smaller range o~ image content values than the dynamic range of the digitized image data base. The code conversion mechanism operates to convert a maximum value of 100% white reflectance to an encoded value that is less than the upper limit of the dynamic range of the database to accommodate shifts in the digitized imagery data and allow for the placement of specular highlights that are beyond the 100% white reflectance maximum.
In accordance with the imagery data processing system described in the above referenced co-pending Kristy application, each high resolution captured image is stored ~s a respective image data file cont~in;n~ a low, or base, resolution image bit m~p file and a plurality of higher resolution residual images associated with respecti~ely increasin~ degrees of image resolution. By iteratively combining these higher resolution residual images with the base resolution image, successively increased resolution images may be recovered from the base resolution image.
As an example, spatial data values representativP of a high resolution (3072 x 2D48) image scan of a 36mm-by- 24mm image frame of a 35mm film strip 10 may be stored as a respective im~ge data file including a base resolution im~ge bit map file containing data values associated with a spatial image array or matrix of 512 rows and 768 columns of pixels and an associated set of residual image files to be , ,. . , ,:
., .. . . . . .. . . .....
-- ~, .
.
-. .. - . - : : ~
.. - :, . :, ; ,:~ : :
W092/0~470 PCT/US9l/06577 ?,~ o-stored on the disc. Within the workstation itself, the base rPsolution image m~y be further sub- sampled to derive an even l~wer resolution sub-array of image values (e.g. o~ the order of 128 x 1~2 pixels) for use by the photofinishing operator in the course of formatting and storin~ a digitized image file.
Pursuant to the present inv~ntion, a substantially reduced resolution version of each di~iti2ed image to be color balance-processed by the scene balance mechanism is also coupled to a scene balance failure discriminator, which determines the likelihood that the output image processed by the scene balance mechanism will be of acceptable color balance quality. For this purpose, the scene balance mechanism failure discriminator P~m;neS the spatial content of a very small sub-array (e.g. 24 x 36 pixels) produced by subsampling compression of a 128 x 192 pixel version of the image of interest and, based upon a previously generated statistical model of prescribed characteristics of the image, derives a measure of the likelihood that the color balance performance of the scene balance mechanism for th~t image will produce a visuall~ acceptable reproduction (high resolution output print).
The discriminator, per se, may comprise a statistically based, linear discriminator, such as a Fischer Discrimin~nt or a Bayes Classifier, which use a multi~;men~ional discrimination space derived from predictors computed from the imagery data. Preferably, however, the failure detection mechanism employs a non-linear discriminator, such as a back propagation-trained neural network, because of its ability to ~5 accurately recognize images that will be correctly .
.. . . .. .
.. . , . - . - - . - --- . ~ :
.. . . ..
.. ... .
, . .
-, , . :. . .,- .
.
2~372~29 processed by the scene balance mechanism. Namely, although both linear discriminators and neural networks are capable of detecting color balance failures for about 50% of the images that will fail ('bad' im~ges), a linear discriminator will also erroneously flag failùres for some images that will be correctly processed by the scene balance mechanism. On the other hand, the preferred neural network-based failure discriminator will effectively never detect a scene balance failure fc)r a 'good' image.
The substantially reduced resolution version of the imaye, such as one comprising â pixel sub-array in the order of only 24x36 pixels, contains sufficient ~5 spatial color content to permit identification of scene balance failures for a high resolution image, but with considerably reduced computational intensity. Still, even with such a reduced image size, the number of values (864 for a 24 x 36 pixel sub-array) that enter into the statistical analysis is still substantial and, for a linear discriminator, may have an associated set of approximately 50 statistical measures (predictors) of the content of the scene. For the same sized image sub- array, it is possible to build a back propagation neural network that operates on a reasonably constrained set of parameters.
As one non-limitative example, a neural network, such as that diay~.u,~tically shown in Figure 2 as a feed-forward n-layer perceptron (trained by back propagation), may be configured of a nine element input layer 31, a twel~e element hidden layer and a single element output layer 33. An image training population on the order of several thousand 24x36 pixel scenes repeatedly and successively applied to the network m~y ~ , . . . .
... .
.: .. . ... - ~ j: . ,... :
-- :: .. . ~ .. .
. ,, ~ , . , . , , .. :. ~:
wos~/0~7~ q~ -12- P~T/US91/06;77 be used to load the network. As a non-limitative example, a set of nine predictors (Nos. l-9) listed in Table l, below and referencing the region map of Figure 6, m~y be employed. It should be observed, however, that other predictors in addition to and different from those listed here, m~y be used. For a discussion of the use of neural networks for pattern recognition and choosing predictors, attention may be directed to the text: ~Adaptive Pattern Recognition and Neural Netwoxks~, by Yoh-Haw Pao, Chapters l, 2 and 5.
T~RT,F~ 1 p~nI~ToR D~FINI~ION
No. l Red maximum of region 5 - red average of entire frame.
No. 2 Red maximums of regions 3, 4, 5 are averaged together and the red average of the entire frame is subtracted from it.
No. 3 Variance of neutral T space value for entire range, where T space is defined, for example, as set forth in U S.
Patents Nos. 4,159,174 and 4,g33,709).
No. 4 Neutral T space average of entire frame.
No. 5 Hyperbolic tangent of center maximumf minus the center ave.age.
No. 6 Neutral T space average of region 5 minus neutral T space average of regions l, 2, 3, 4.
N0. 7 Average of neutral T space maximums of regions 3, 4, 5 minus neutral T space average of entire frame.
No. 8 Neutral T space centroid of entire frame.
No. 9 Green magenta average of entire frame.
.. .. - , - - - -. ,:
W0~2/0~7~ PCT/US91/06577 -13- 2~72~2~
Moreover, the type of neural network employed - is not constrained to a feed-forward, back propagation implementation. Other network types, such as a counter propagation network may be used. Suffice it to say that examples given here have been shown to correctly identify scene balance failures for approximately 50%
of 'bad' images (ima~es for which the scene balance mechanism will fail) but 0~ of 'good' images, namely those images for which the scene balance mechanism will produce a properly color-balanced output image. For a further discussion of neural networks in general and their application to statistica} processing, attention m~y be directed to the text: ~Parallel Distributed Processing~ Vol. I-II, by McClelland, Rumelhart and PDP
Research Group, MIT Press, 1988.
The scene balance failure detection mechanism of the present invention is preferably incorporated within the image processing software employed by workstation 14, in order that those images where the scene balance mechanism is likely to fail will be identified and corrected prior to the reproduction (printing) stage of the photofinishing process. Once such a scene balance failure image is identified, one or more reasonably low resolution versions of the digitized image processed by the scene balance mech~nism are displayed to the photofinishing operator, tog~ ner with a request for the photofinishing operator 3Q to input image adjustment information that is used to modify or correct the color balance action of the scene balance mechanism on that im~ge. In response to this user input information, ~he image processor adjusts color balance parameters of the scene balance mechanism, so that the resulting processed image will . . .
: , , - -: - . , : , .", ., . :: .: : :. . ; .
: . .. , . -, ~ , :. :
WVg/0;47~ PC~/US~110677 produce a print of acceptable color balance quality.
This interaction of the photofinishing operator with the image processing system may be accomplished using any one of several modes dia~.rammatically illus~rated in Figure 3-5, described;below.
More particulaxly, Figure 3 illustrates a first mode of the failure detection and interactive correction process according to the present invention in which a plurality of low resolution (128 x 192 pixels) images 20-l, 20-2, 20-3,..., 20-N, each of which is a respectively different version of the (l28 x 192) digitized image a~ processed by the resident scene balance mechanism, are displayed as a linear array on the photofinisher's workstation display. The image tha~
results from the application of the currently defined scene balance mechanism to the scanned image may correspond to image 20-3, centrally placed within a linear array of images across the workstation display screen. Images 20-l, 20-2 and im~ges 20-4 and 20-5 on opposite sides of this image have respectively different variations in color balance obtained by successively offsetting or shifting the scene balance mechanism along a prescribed axis (e.g. lighter and darker scan densities where the failure mode detected is one of neutral balance~. Further variations in the image may be formed by the use of a conventional graphics processing scrolling operator to effectively create a 'sliding scale' for a linear array of images that may be panned across the screen.
By way of a display prompt, the photofinishing operator is requested to choose the most visually pleasing one of the different versions of the image within the display window associated with .
- . -: .
- - : . . , ~~ -15- ~ ~ 2 ~ 2~
respecti~ely different biasing of the sc-ne balance mechanism. In response to this selection, the manner in which the scene balance mechanism processes the digiti~ed image is modified to conform with the color - S balance characteristics of the selected image version.
For this purpose, the bias settings employed to produce the selected image are used to modify the scene balance mechanism. Namely, since each image version represents a prescribed offse~ or shift of th~ scene balance mechanism, selecting a given image effectively ' specifies the magnitude of the offset bias to be used by the scene balance mechanism in processing the high resolution image for storage and subsequent reproduction.
~5 Figure 4 illustrates a second mode of the failure detection and interactive correction process according to the present invention in which the photofinishing minilab operator is given a choice of only two versions of the image. As shown in the figure, on the workstation display two low resolution images 41 and 42, each of which is a respectively dif erent version of the (128 x 192) digitized image as processed by the resident scene balance mechanism, are displayed on the photofinisher's workstation display. One of the displayed images has been processed as though the image is front-lit (e.g. by a flash illumination source), while the other displayed ima~e has been processed as though the image is back-lit ~silhouette). Both types of images look similar to the scene balance mechanism, but -equire different processing. The photofinisher~s choice of which displayed version of the image is most pleasing determines which version of the scene balance mechanism will be used to process that image.
.. . . . . . ...
. ~ :
:, , -w~92/0~70 PCT/US91/06~77 Figure 5 diagrammaticallY illustrates a third mode of operation of the scene balance failure discriminator, whereln a low resolution version 51 of the image processed by the scene balance mechanism is displayed to the photofinishing operator, together with a panel 52 containing a question regarding the type of image, e.g. whether cr not the image is backlit or a flash scene. The photofinisher's input response to the displayed que~tion panel 52 ~e.g. by mouse or keyboard entry) is then employed by the image processing software to refine how the scene balance mechanism corrects the image (e.g. corrects ~or an ambiguity in the type of scene illumination). For example, for the case where the photofinisher enters the fact that the subject/scene is backlit, a scene balance mechanism optimized for a silhouette is employed. Similarly, for the case where the photofinisher enters the fact that the subject/scene is illuminated by a flash source, a scene balance mechanism optimized for a front-lit image is used.
As will be appreciated from the foregoing description, the present invention provides a technique for detecting images prone to scene balance failures and enabling the photofinisher to interactively correct for failures in the perform~nce of ~he scPne balance mechanism before an image has been processed and an unacceptable output im2ge printed. By correcting for a discovered scene balance failure prior to printing the image, the photofinisher will not incur what can amount to substantial penalties in terms of cost of materials and film processing time.
Moreover, the present invention offers the advantage that, by using a failure discriminator, only .
,- :
.:
: . ,-: ' --- :, . .
. ..
,, :.... : --wn g~/05470 PCT/US9l/~6577 2~262~
some, but not all, of the scenes need to by evaluated by the photofinisher. Without it, to detect as many failures as possible, the photofinisher would have to look at every im~ge.
While we have shown and described an embodiment in accordance with the present invention, it is to be understood that the same is not limited thereto but is susceptible to numerous changes and modifications as known to a person skilled in the art, and we therefore do not wish to be limited to the details shown and described herein but intend to cover all such changes and modifications as are obvious to one of ordinary skill in the art.
.. ..
_ . . . .. . . . . ... . . . . . . . . .. ..
, :: .: ' . ' :, . r
Claims (28)
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:
1. For use with an image processing system, in which a color photographic image that has been captured on a color photographic image recording medium is digitally encoded by an opto-electronic device and stored in a digital data base as a digitized image, said digitized image being processed by means of an image adjustment mechanism to derive a modified output image to be coupled to an image reproduction device, a method of controlling the operation of said system to reduce the effects of failures in the processing of a digitized image by said image adjustment mechanism comprising the steps of:
(a) processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment mechanism will cause said output image to be reproduced as a prescribed quality image; and (b) in response to a prescribed likelihood that said image adjustment mechanism will fail to cause said output image to be reproduced as said prescribed quality image, generating a request for user-sourced control information for modifying the manner in which said image adjustment mechanism processes said image.
(a) processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment mechanism will cause said output image to be reproduced as a prescribed quality image; and (b) in response to a prescribed likelihood that said image adjustment mechanism will fail to cause said output image to be reproduced as said prescribed quality image, generating a request for user-sourced control information for modifying the manner in which said image adjustment mechanism processes said image.
2. A method according to claim 1, wherein step (b) comprises, in response to a prescribed likelihood that said image adjustment mechanism will fail to cause said image to be reproduced as said prescribed quality image, displaying an image representative of the effect of the processing of said image by said image adjustment mechanism and generating a request for user-sourced control information for controllably modifying the manner in which said image adjustment mechanism processes said image.
3. A method according to claim 1, wherein step (b) comprises, in response to a prescribed likelihood that said image adjustment mechanism will fail to cause said image to be reproduced as said prescribed quality image, displaying a plurality of images representative of the effects of different processing of said image by modified versions of said image adjustment mechanism and generating a request for a user-sourced selection of one of said displayed images, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction.
4. A method according to claim 1, wherein said image adjustment mechanism comprises a scene balance mechanism which processes a digitized image to derive a color content-balanced output image.
5. A method according to claim 1, wherein said image adjustment mechanism failure discriminator is implemented as a statistical linear discriminator.
6. A method according to claim 1, wherein said image adjustment mechanism failure discriminator is implemented as a non-linear discriminator.
7. A method according to claim 1, wherein said image adjustment mechanism failure discriminator is implemented as a neural network.
8. A method according to claim 7, wherein said neural network is a back propagation-trained neural network.
9. A method according to claim 1, wherein step (b) comprises displaying an image representative of the effect of processing said digitized image by said image adjustment mechanism and generating a request for the inputting of information representative of a characteristic of said image, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction.
10. A method according to claim 1, wherein step (a) comprises coupling a reduced resolution version of said digitized image to said image adjustment mechanism failure discriminator in order to define the likelihood of whether or not said image adjustment mechanism will cause said output image to be reproduced as a prescribed quality image.
11. For use with an image processing system, in which a color photographic image that has been captured on a color photographic image recording medium is digitally encoded by an opto-electronic device and stored in a digital data base as a digitized image, said digitized image being processed by means of an image adjustment mechanism to derive a modified output image to be coupled to an image reproduction device, a method of controlling the operation of said system to reduce the effects of failures in the processing of a digitized image by said image adjustment mechanism comprising the step of:
(a) prior to reproducing said image, processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment mechanism will cause said output image to be reproduced as a prescribed quality image.
(a) prior to reproducing said image, processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment mechanism will cause said output image to be reproduced as a prescribed quality image.
12. A method according to claim 11, further including the step of:
(b) in response to said image adjustment failure discriminator indicating a prescribed likelihood that said image adjustment mechanism will fail to cause said output image to be reproduced as said prescribed quality image, modifying said image adjustment mechanism.
(b) in response to said image adjustment failure discriminator indicating a prescribed likelihood that said image adjustment mechanism will fail to cause said output image to be reproduced as said prescribed quality image, modifying said image adjustment mechanism.
13. A method according to claim 12, wherein step (b) comprises, in response to said image adjustment failure discriminator indicating a prescribed likelihood that said image adjustment mechanism will fail to cause said image to be reproduced as said prescribed quality image, displaying an image representative of the effect of the processing of said digitized image by said image adjustment mechanism and generating a request for control information for controllably modifying the manner in which said image adjustment mechanism processes said digitized image.
14. A method according to claim 12, wherein step (b) comprises, in response to said image adjustment failure discriminator indicating a prescribed likelihood that said image adjustment mechanism will fail to cause said image to be reproduced as said prescribed quality image, displaying a plurality of images representative of the effects of different processing of said digitized image by modified versions of said image adjustment mechanism and generating a request for a selection of one of said displayed images, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction.
15. A method according to claim 11, wherein said image adjustment mechanism comprises a scene balance mechanism which processes an image to derive a color content-balanced output image.
16. A method according to claim 11, wherein said image adjustment mechanism failure discriminator is implemented as a statistical linear discriminator.
17. A method according to claim 11, wherein said image adjustment mechanism failure discriminator is implemented as a non-linear discriminator.
18. A method according to claim 15, wherein step (a) comprises coupling the contents of said digitized image to a neural network-implemented scene balance mechanism failure discriminator.
19. A method according to claim 18, wherein said neural network is a back propagation-trained neural network.
20. A method according to claim 11, wherein step (b) comprises displaying an image representative of the effect of processing said digitized image by said image adjustment mechanism and generating a request for the inputting of information representative of a characteristic of said image, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction.
21. A method according to claim 11, wherein step (a) comprises coupling a reduced resolution version of said digitized image to said image adjustment mechanism failure discriminator in order to define the likelihood of whether or not said image adjustment mechanism will cause said output image to be reproduced as a prescribed quality image.
22. For use with an image processing system, in which a color photographic image that has been captured on a color photographic image recording medium is digitally encoded by an opto-electronic device and stored in a digital data base as a digitized image, said digitized image being processed by means of an image adjustment mechanism to derive a modified output image to be coupled to an image reproduction device, a method of controlling the operation of said system to reduce the effects of failures in the processing of a digitized image by said image adjustment mechanism comprising the step of:
displaying an image representative of the effect of the processing of said digitized image by said image adjustment mechanism and generating a request for control information for controllably modifying the manner in which said image adjustment mechanism processes said digitized image, and wherein said step comprises displaying a plurality of images representative of the effects of different processing of said digitized image by modified versions of said image adjustment mechanism and generating a request for a selection of one of said displayed images, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction; and wherein said step includes the preliminary step of processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment failure discriminator will cause said modified output image to be reproduced by said image reproduction device as a prescribed quality image.
displaying an image representative of the effect of the processing of said digitized image by said image adjustment mechanism and generating a request for control information for controllably modifying the manner in which said image adjustment mechanism processes said digitized image, and wherein said step comprises displaying a plurality of images representative of the effects of different processing of said digitized image by modified versions of said image adjustment mechanism and generating a request for a selection of one of said displayed images, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction; and wherein said step includes the preliminary step of processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment failure discriminator will cause said modified output image to be reproduced by said image reproduction device as a prescribed quality image.
23. For use with an image processing system, in which a color photographic image that has been captured on a color photographic image recording medium is digitally encoded by an opto-electronic device and stored in a digital data base as a digitized image, said digitized image being processed by means of an image adjustment mechanism to derive a modified output image to be coupled to an image reproduction device, a method of controlling the operation of said system to reduce the effects of failures in the processing of a digitized image by said image adjustment mechanism comprising the step of:
displaying an image representative of the effect of the processing of said digitized image by said image adjustment mechanism and generating a request for control information for controllably modifying the manner in which said image adjustment mechanism processes said digitized image, and wherein said step comprises displaying an image representative of the effect of processing said digitized image by said image adjustment mechanism and generating a request for the inputting of information representative of a characteristic of said image, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction; and wherein said step of includes the preliminary step of processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment failure discriminator will cause said modified output image to be reproduced by said image reproduction device as a prescribed quality image.
displaying an image representative of the effect of the processing of said digitized image by said image adjustment mechanism and generating a request for control information for controllably modifying the manner in which said image adjustment mechanism processes said digitized image, and wherein said step comprises displaying an image representative of the effect of processing said digitized image by said image adjustment mechanism and generating a request for the inputting of information representative of a characteristic of said image, in response to which said image adjustment mechanism is controllably adjusted to process said image as an acceptable quality output image for reproduction; and wherein said step of includes the preliminary step of processing said digitized image in accordance with an image adjustment failure discriminator to define the likelihood of whether or not said image adjustment failure discriminator will cause said modified output image to be reproduced by said image reproduction device as a prescribed quality image.
24. A method according to claim 22, wherein said image adjustment mechanism comprises a scene balance mechanism which processes a digitized image to derive a color content-balanced output image.
25. A method according to claim 22, wherein said image adjustment mechanism failure discriminator is implemented as a statistical linear discriminator
26. A method according to claim 22, wherein said image adjustment mechanism failure discriminator is implemented as a non-linear discriminator.
27. A method according to claim 22, wherein said image adjustment mechanism failure discriminator is implemented as a neural network.
28. A method according to claim 27, wherein said neural network is a back propagation-trained neural network.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US583,740 | 1990-09-17 | ||
| US07/583,740 US5644647A (en) | 1990-09-17 | 1990-09-17 | User-interactive reduction of scene balance failures |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CA2072629A1 CA2072629A1 (en) | 1992-03-18 |
| CA2072629C true CA2072629C (en) | 1997-12-16 |
Family
ID=24334367
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002072629A Expired - Fee Related CA2072629C (en) | 1990-09-17 | 1991-09-13 | User-interactive reduction of scene balance failures |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US5644647A (en) |
| EP (1) | EP0500926B1 (en) |
| JP (1) | JPH05502741A (en) |
| CA (1) | CA2072629C (en) |
| DE (1) | DE69125700T2 (en) |
| WO (1) | WO1992005470A1 (en) |
Families Citing this family (39)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5354199A (en) * | 1991-08-02 | 1994-10-11 | Minnesota Mining And Manufacturing Company | Adhesive for packaged orthodontic appliance |
| US5748342A (en) * | 1994-04-18 | 1998-05-05 | Canon Kabushiki Kaisha | Image processing apparatus and method |
| WO1996009717A1 (en) * | 1994-09-19 | 1996-03-28 | Apple Computer, Inc. | Generation of tone reproduction curves using psychophysical data |
| US5884296A (en) * | 1995-03-13 | 1999-03-16 | Minolta Co., Ltd. | Network and image area attribute discriminating device and method for use with said neural network |
| WO1996035291A1 (en) * | 1995-05-02 | 1996-11-07 | Polaroid Corporation | Apparatus for use in color correction in an electronic imaging system |
| US6786420B1 (en) | 1997-07-15 | 2004-09-07 | Silverbrook Research Pty. Ltd. | Data distribution mechanism in the form of ink dots on cards |
| CN1167265C (en) | 1997-06-30 | 2004-09-15 | 富士写真胶片株式会社 | Image communication system and method thereof |
| JPH1132215A (en) * | 1997-07-10 | 1999-02-02 | Fuji Photo Film Co Ltd | Photographic image, method and device for forming it, and recording medium therefor |
| US6618117B2 (en) * | 1997-07-12 | 2003-09-09 | Silverbrook Research Pty Ltd | Image sensing apparatus including a microcontroller |
| US7551201B2 (en) | 1997-07-15 | 2009-06-23 | Silverbrook Research Pty Ltd | Image capture and processing device for a print on demand digital camera system |
| US6879341B1 (en) | 1997-07-15 | 2005-04-12 | Silverbrook Research Pty Ltd | Digital camera system containing a VLIW vector processor |
| US6624848B1 (en) * | 1997-07-15 | 2003-09-23 | Silverbrook Research Pty Ltd | Cascading image modification using multiple digital cameras incorporating image processing |
| US6690419B1 (en) | 1997-07-15 | 2004-02-10 | Silverbrook Research Pty Ltd | Utilising eye detection methods for image processing in a digital image camera |
| US7110024B1 (en) | 1997-07-15 | 2006-09-19 | Silverbrook Research Pty Ltd | Digital camera system having motion deblurring means |
| US6101000A (en) * | 1998-01-30 | 2000-08-08 | Eastman Kodak Company | Photographic processing apparatus and method |
| US6222613B1 (en) * | 1998-02-10 | 2001-04-24 | Konica Corporation | Image processing method and apparatus |
| AUPP702098A0 (en) | 1998-11-09 | 1998-12-03 | Silverbrook Research Pty Ltd | Image creation method and apparatus (ART73) |
| US6208770B1 (en) * | 1998-09-18 | 2001-03-27 | Eastman Kodak Company | Digital colored corrected prints produced from colored film |
| US6263792B1 (en) * | 1999-02-12 | 2001-07-24 | Eastman Kodak Company | Method and apparatus for printing digital images |
| JP3903083B2 (en) * | 1999-03-11 | 2007-04-11 | 富士フイルム株式会社 | Image processing apparatus, method, and recording medium |
| US6781713B1 (en) * | 1999-05-20 | 2004-08-24 | Eastman Kodak Company | Correcting exposure in a rendered digital image |
| AUPQ056099A0 (en) | 1999-05-25 | 1999-06-17 | Silverbrook Research Pty Ltd | A method and apparatus (pprint01) |
| US6581109B1 (en) | 1999-06-24 | 2003-06-17 | International Business Machines Corporation | System to dynamically adjust image colors in client/server environment to assure accurate color reproduction |
| US20020085743A1 (en) * | 2000-04-04 | 2002-07-04 | Konica Corporation | Image processing selecting method, image selecting method and image processing apparatus |
| US6490023B1 (en) | 2000-06-13 | 2002-12-03 | Eastman Kodak Company | Processing apparatus and method for determining optimum processing characteristics of thermal developable film |
| US6781724B1 (en) * | 2000-06-13 | 2004-08-24 | Eastman Kodak Company | Image processing and manipulation system |
| US6369873B1 (en) | 2000-06-13 | 2002-04-09 | Eastman Kodak Company | Thermal processing system and method including a kiosk |
| US7023448B1 (en) * | 2000-08-22 | 2006-04-04 | Adobe Systems Incorporated | Selecting rendering intent |
| US6985637B1 (en) * | 2000-11-10 | 2006-01-10 | Eastman Kodak Company | Method and apparatus of enhancing a digital image using multiple selected digital images |
| WO2002078344A1 (en) * | 2001-03-26 | 2002-10-03 | Hyun-Gi An | Method for processing moving pictures responding to user's action |
| US6999615B2 (en) | 2001-06-05 | 2006-02-14 | Basf Corporation | Color management and solution distribution system and method |
| US20020184168A1 (en) * | 2001-06-05 | 2002-12-05 | Mcclanahan Craig J. | System and method for determining acceptability of proposed color solution using an artificial intelligence based tolerance model |
| US6993512B2 (en) | 2001-06-05 | 2006-01-31 | Basf Corporation | System and method for converting a color formula using an artificial intelligence based conversion model |
| US7054484B2 (en) * | 2001-07-14 | 2006-05-30 | Lyford Kirk S | Guided color correction system |
| FR2858088B1 (en) * | 2003-07-25 | 2005-10-07 | Eastman Kodak Co | METHOD FOR DIGITAL MODIFICATION OF IMAGE EXPOSURE |
| US20050219561A1 (en) * | 2003-11-28 | 2005-10-06 | Canon Kabushiki Kaisha | User interface for selecting color settings |
| JP4338551B2 (en) * | 2004-03-04 | 2009-10-07 | 富士フイルム株式会社 | Image reproduction method, image reproduction apparatus, and image reproduction program |
| JP4327754B2 (en) | 2005-03-16 | 2009-09-09 | 富士フイルム株式会社 | Image list generation apparatus and method, and program |
| US7830428B2 (en) | 2007-04-12 | 2010-11-09 | Aptina Imaging Corporation | Method, apparatus and system providing green-green imbalance compensation |
Family Cites Families (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CH651403A5 (en) * | 1980-12-19 | 1985-09-13 | Walter Dr Grossmann | Method and device for exposure of recording equipment in the production photographic colour copies. |
| DE3205898A1 (en) * | 1982-02-18 | 1983-08-25 | Agfa-Gevaert Ag, 5090 Leverkusen | METHOD AND DEVICE FOR PREPARING PHOTOGRAPHIC TEMPLATE STRIPS FOR COPYING |
| DE3437069A1 (en) * | 1984-10-09 | 1986-04-10 | Agfa-Gevaert Ag, 5090 Leverkusen | METHOD AND DEVICE FOR PRODUCING COPIES OF COLOR DOCUMENTS DIFFICULT TO COPY |
| US4685071A (en) * | 1985-03-18 | 1987-08-04 | Eastman Kodak Company | Method for determining the color of a scene illuminant from a color image |
| US4642684A (en) * | 1985-05-06 | 1987-02-10 | Eastman Kodak Company | Method and apparatus for detecting and printing color photographic images of scenes exposed with narrow band illuminant |
| JPS61275625A (en) * | 1985-05-31 | 1986-12-05 | Fuji Photo Film Co Ltd | Calibrating method for color photographic image information |
| US4677465A (en) * | 1985-11-01 | 1987-06-30 | Eastman Kodak Company | Digital color image processing method with shape correction of histograms used to produce color reproduction functions |
| JPS62141530A (en) * | 1985-12-16 | 1987-06-25 | Fuji Photo Film Co Ltd | Color negative testing device |
| US4953227A (en) * | 1986-01-31 | 1990-08-28 | Canon Kabushiki Kaisha | Image mosaic-processing method and apparatus |
| JPS62189456A (en) * | 1986-02-17 | 1987-08-19 | Fuji Photo Film Co Ltd | Main picture detection method and photographic printing exposure decision method employing said detection |
| JPH0640198B2 (en) * | 1986-02-17 | 1994-05-25 | 富士写真フイルム株式会社 | Method for detecting main image and method for determining exposure amount for photo printing using the same |
| JPH0675161B2 (en) * | 1986-06-13 | 1994-09-21 | 富士写真フイルム株式会社 | Photographic printing condition inspection device |
| US4942462A (en) * | 1987-07-15 | 1990-07-17 | Fuji Photo Film Co., Ltd. | Photographic printer having a CRT for illuminating an original with a flying spot to print an image of the original |
| US4899212A (en) * | 1987-08-28 | 1990-02-06 | Fuji Photo Film Co., Ltd. | White balance adjusting device for a camera |
| DE68916049T2 (en) * | 1988-02-05 | 1994-09-22 | Fuji Photo Film Co Ltd | Method and apparatus for viewing color photographic films. |
| US4979031A (en) * | 1988-07-12 | 1990-12-18 | Minolta Camera Kabushiki Kaisha | Digital color copying machine |
-
1990
- 1990-09-17 US US07/583,740 patent/US5644647A/en not_active Expired - Lifetime
-
1991
- 1991-09-13 DE DE69125700T patent/DE69125700T2/en not_active Expired - Lifetime
- 1991-09-13 EP EP91918969A patent/EP0500926B1/en not_active Expired - Lifetime
- 1991-09-13 WO PCT/US1991/006577 patent/WO1992005470A1/en active IP Right Grant
- 1991-09-13 CA CA002072629A patent/CA2072629C/en not_active Expired - Fee Related
- 1991-09-13 JP JP3517251A patent/JPH05502741A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| CA2072629A1 (en) | 1992-03-18 |
| US5644647A (en) | 1997-07-01 |
| DE69125700D1 (en) | 1997-05-22 |
| EP0500926A1 (en) | 1992-09-02 |
| DE69125700T2 (en) | 1997-10-23 |
| WO1992005470A1 (en) | 1992-04-02 |
| JPH05502741A (en) | 1993-05-13 |
| EP0500926B1 (en) | 1997-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2072629C (en) | User-interactive reduction of scene balance failures | |
| US5371615A (en) | Image-dependent color correction using black point and white point in a natural scene pictorial image | |
| US7006688B2 (en) | Histogram adjustment features for use in imaging technologies | |
| Mann | Comparametric equations with practical applications in quantigraphic image processing | |
| US5528339A (en) | Color image reproduction of scenes with color enhancement and preferential tone mapping | |
| US7260323B2 (en) | Imaging using silver halide films with micro-lens capture, scanning and digital reconstruction | |
| EP0961484B1 (en) | Digital photofinishing system with digital image processing | |
| EP0221929B1 (en) | Digital color image processing employing histogram normalization | |
| EP0589376B1 (en) | Colour image reproduction of scenes with preferential tone mapping | |
| US6304345B1 (en) | Auto resoration of a print | |
| Schreiber | Image processing for quality improvement | |
| US7034878B2 (en) | Camera apparatus and method of taking pictures including simplified processing | |
| US5796874A (en) | Restoration of faded images | |
| EP0878777A2 (en) | Method for enhancement of reduced color set images | |
| US6603886B1 (en) | Image processing method and system | |
| US5282063A (en) | Method and apparatus for automatic brightness and contrast control in an image capture system | |
| JP2710310B2 (en) | Electronically printing a copy of a painting manuscript | |
| US6459500B1 (en) | Image processing apparatus | |
| JPH02118642A (en) | Color correction by dry dot etching using mask produced photographically | |
| JP3408770B2 (en) | Image processing device | |
| US6433888B1 (en) | Auto restoration of a print | |
| JP2001222710A (en) | Device and method for image processing | |
| US20030011790A1 (en) | Correcting exposure and tone scale of digital images using a plurality of transforms | |
| JPH07121681A (en) | Automatic color tone corrector | |
| WO2001078368A2 (en) | Film and video bi-directional color matching system and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| MKLA | Lapsed |