CA1213989A - Multiple measurement noise reducing system using space-variant filters - Google Patents
Multiple measurement noise reducing system using space-variant filtersInfo
- Publication number
- CA1213989A CA1213989A CA000447768A CA447768A CA1213989A CA 1213989 A CA1213989 A CA 1213989A CA 000447768 A CA000447768 A CA 000447768A CA 447768 A CA447768 A CA 447768A CA 1213989 A CA1213989 A CA 1213989A
- Authority
- CA
- Canada
- Prior art keywords
- image signal
- processed image
- measurements
- signal
- pass filtering
- 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
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/30—Transforming light or analogous information into electric information
- H04N5/32—Transforming X-rays
Abstract
ABSTRACT
Extraneous signals or artifacts are reduced in a multiple measurement noise reducing system such as an x-ray imaging system by processing a plurality of measurements to obtain a first image signal of an object representing a desired parameter such as a blood vessel, processing the plurality of measurements to provide a second image signal having increased signal-to-noise, low pass filtering the first image signal, high pass filtering the second image signal, and then combining the two filtered signals. The filter frequencies are varied in response to the presence of arti-facts to minimize effects of the artifact on the combined signal.
Extraneous signals or artifacts are reduced in a multiple measurement noise reducing system such as an x-ray imaging system by processing a plurality of measurements to obtain a first image signal of an object representing a desired parameter such as a blood vessel, processing the plurality of measurements to provide a second image signal having increased signal-to-noise, low pass filtering the first image signal, high pass filtering the second image signal, and then combining the two filtered signals. The filter frequencies are varied in response to the presence of arti-facts to minimize effects of the artifact on the combined signal.
Description
A-38870-'~KW
~J~ 3~ ~
MULTIPLE MEASUREMENT NOISE RED~CING SYS~EM USING SPACE-VARIANT FILTERS
This invention relates generally tO imaging systems such as x-ray systems, and more particularly the invention relates to the processing of images derived from a plurality of measurements.
In many imaging application areas, images are constructed as a weighted sum of a plurality of measurements. A prime example are the recent new developments in x-ray imaging.
Here measurements are made at different energies and/or at different times. These measurements are then combined to provide selective images representing specific materials.
One example is that of temporal subtraction techniques using digital radiography as described in the publication ~y C. A.
Mistretta and A. B. Crummy, "Diagnosis of Cardiovascular Disease by Digital Subtraction Angiography," in Science, Vol. 214, pp. 761-65, 1981. Here measurements taken before and after the administration of iodine into blood vessels are subtracted to provide an image of vessels alone. Another example is that of energy-selective radiography as described in the publication by L. A. Lehmann, et al., "Generalized Image Combinations in Dual KVP Digital Radiography," in Medical Physics, Vol. 8, pp. 659-67, 1931. Here measure-ments made at different energies are combined to enhance or eliminate specific materials. A third example is energy-selective computerized tomography as described in the ~L~ 3~
publication by R. E. Alvarez and A. Macovski, " Energy-Selec-tive Reconstructions in X--ray Computerized Tomography," in Physics in Medicine ~ Bioloqyr Vol. 21, pp. 733-744, 1976. Here sets of measurements are made at two energy spectra and distortion-free reconstructions are made which delinea-te the average atomic number and densi-ty of each region. A fourth e~ample is hybrid subtraction described in the publication by W. R. Brody, "Hybrid Subtraction for Improved Intravenous Arteriography," in Radioloqy, Vol. 141, pp. 828-831, 1981. Here dual energy measurements are 1() made before and after the administration of iodine. Each dual energy pair is used to elimi,nate soft tissue. In this way the resultant subtracted vessel images are immune to soft tissue motion.
In each case, where a number of measurements are combined to select specific material properties, the resultant SNR (Signal-to-noise ratio~ is reduced as compared to the non-selective image.
Therefore, these various processing techniques which improve the visualization of disease processes by selecting specific materials have the disadvantage of a reduc-tion in SNR. This reduced SNR can interfere with the ability to visualize regions of disease.
A system to improve the SNR of the hybrid subtraction and other multiple measurement systems has been developed by A. Macovski. An array of measurements are taken of an object under different conditions. These are combined to select a specific aspect of the object. The same measuremen-ts are then combined in different weightings to provide a lower-noise image, wi-thout the desired selectivity. The improved selective image is formed by combining the selective image with the lower-noise image - 2a - ~ 9 using the lower frequency components of ~he former and the higher frequency components of the latter. To assure the proper amplitude of the higher frequency componen-ts, the components are weighted such as with the ratio of the derlvative of the selective image to the derivative of the lower-noise image.
Nonlinear thresh~lds are used for those cases where this ratio becomes an unreliable indicator of the correct amplitude of the high frequency components. One limitation of tne Macovski system is the introduction of extraneous signals (herein termed "artifacts"3 by the low noise signal.
The present invention is directed to an improvement in the Macovski system by minimizing the artifacts introduced in the displayed image. In accordance with the invention, a space-variant filter is applied to a selective image signal, S, and the complementary space-variant filter to a lower noise image signal, M. The sum of these two filtered images appropriately scaled results in the improved image S. This process still combines the low frequency components of S
wi~h the high frequency components of M, but the spatial frequency distribution of the filter response varies with the picture element. The procedure may be expressed by S = Sl(*) + Mh(*) where Sl(*) and Mh(*) represent the two filtered images, l(*) the space-variant filter, and h(*) the complementary space-variant filter. Because the two filters are always complementary and the combination is linear, an important property is that the desired signal component is never distorted in the processed image. In the origina~ NRS, distortion is possible in the form of blurring when the weighting of the high pass filtered M image ~Mh), which contains the edge information, becomes small to prevent artifact edges from entering the enhanced image. In the improved system, the filter used at each picture element is chosen to suppress the amount of high frequency artifact that leaks into the processed image without undue sacrifice of the SNR.
Accordingly, an object of the present invention is a multiple measurement noise reducing system which minimizes e~traneous ~:1 3~
artifacts.
Another object of -the inventi.on is a method of improving the image in a mul.tiple measurement noise reducing system.
According to one aspect of the invention there is provided for use in an imaging system, a method of reducing extraneous signals in a first processed image signal of an object representing a desired parameter derived from a plurali-ty of measurements comprising the steps of: processing said plurali-ty of measurements to provide a second processed image signal having greater signal--to-noise ratio than said first processed signal, low pass filtering said first processed image signal to reduce its noise, high pass filtering said second processed image signal, said high pass filtering being complementary to said low pass filtering; varying the filter frequencies of said low pass fiitering and said high pass filtering in response to presence of artifacts, and combining said filtered first processed image and said filtered second processed image to provide an improved processed image signal.
Also, according to the invention, there is provided in an imaging system, apparatus for reducing ex-traneous image signals in a processed image signal of an object representing a desired parameter derived from a plurality of measurements comprising:
first processing means for processing said plurality of measure-ments and producing a first processed image signal representing said desired parameter, second processing means for processing said plurality of measurements and producing a second processed image signal having a greater signal-to-noise ratio than said firs-t processed image signal, first fil-ter means for receiving - 4a -and low pass filtering said first processed image signal, second filter means for receiving and high pass filtering said second processed imaye signal, means for varying the filter band widths of said low pass filter means and said high pass filter means in response to presence of extraneous image signals, and means for combining said filtered first processed image signal and said filtered second processed image signal.
The inven-tion and objects and features thereof will be more readily apparent from the following detailed description and appended claims when taking with the drawing, in which:
Figure 1 is a block diagram of an x-ray system for acquiring signals useful in the invention.
Figure 2 is a block diagram of an x-ray system in accordance with -the prior art.
Figure 3 is a functional block diagram of an x-ray system in accordance with the present invention.
Figure 4 is a functional block diagram of edge detection circuitry useful with the system of Figure 3.
Referring now to the drawings, Figure 1 is a block diagram of an x-ray system which includes an x-ray tube 8 supplied by variable voltage power source 3. The x-rays are transmitted through object 1, normally the human anatomy. The transmitted rays are received by x-ray detector 4, such as an image intensi-fier and television camera, or a one- or two-dimensional array of scintillators, etc. The image signal is applied to storage system 6 where a plurality of images 9, 10, 11, 12, etc., can be stored.
In temporal sub-traction a first image 9 is s-tored. Follow-~ 3 ~7~ 9 ing administration of a contrast agen-t using syringe 2, a second image 10 is stored. These are -then used in the subsequent processing systems. In energy selec-tive systems, images 9, 10, 11, 12, e-tc., are stored a-t different values of beam A-38870~HKW
-5- ~ 3~5~
energy corresponding to different anode voltages 3. In dual-energy systems, two voltages are used. In addition different x-ray beam filters, not shown, can be ad~ed.
In the Brody hybrid subtraction s~stem previously described, measurements at two voltages are taken prior to the admini-stration of contrast material. These are stored as 9 and 10. Following administration of the contrast agent, two additional measurements at the same two voltages are taken and stored as 11 and 12. ThesP four measurements are then processed to obtain a low-noise image of the vessels opacified by the contrast material from syringe 2.
Referring now to ~igure 2, the lines 9, 10, 11 and 12 repre-senting an array of x-ray measurements of an object, are processed in accordance with the Macovski system described above. These signals could be obtained, for example, from the output of the television camera in a fluoroscopic system where an array of images are formed and stored, for example, on a magnetic disc or in a solid-state frame memory. These signals can represent x-ray images taken with different photon energy spectra or taken before and after the admini-stration of a contrast agent.
In general the multiple measurements are taken so as to provide the isolation or elimination of specific materials.
One simple example involves the isolation of vessel images by subtracting measurements taken before and after the administration of an iodinated contrast agent. Other examples include measurements at different x-ray photon ener~y spectra which are then combined to eliminate or enhance specific materials.
Weighted summer 13 takes the measurements and provides selective signal 17, containing the desired material pro-perties. Unfortunately, however, signal 17 often has relatively poor SNR. The weighted sum usuall~v involves negative weights which diminish the signal. The noise variances of each independent measurement add, however, to A-3887~)-HKW
-6~
provide a large ~ariance, hence a low SNR. This SNR is considerably improved using low pass filter 15 which reduces the bandwidth. This noise reduction, however, i5 accom-plished at the price of reduced resolution. A primary objective of this invention is to restore the resolution using a source of high frequency information with reduced noise.
Signals 9, 10, 11 and 12 are also applied to weighted summing structure 14. These weights can be chosen to maximize the signal-to-noise ratio for a particular component of the image. For example, for vessel imaging they can be chosen to maximize the sisnal-to-noise ratio for the iodinated contrast agent. This is in sharp distinction to weighted summing structure 13 where the weights are chosen to isolate iodine and eliminate the various tissue components. Ideally, the weights which ma~imize the SNR will vary in different portions of the image, depending on the intervening tissue.
Thus, the weights can be varied as the signals axe scanned.
However, for simplicity, the weights can be fixed to maximize the average SNR with relatively small degradation in per-formance.
The high SNR weighted-sum signal, 18, is applied to high pass filter 16 which extracts the high frequency components.
A preferred embodiment of this high pass filter is the complement of low pass filter 15. Thus the sum of the normalized transmissions equals unity within the frequencies of interest. These filters can either be implemented in the frequency domain or as convolution functions.
Combiner 21 combines the low frequency image or signal 19 with the high frequency image or slgnal 20 to obtain the processed image 22 which is displayed in display 23. Since signal 19 is low noise because of the filtering, and signal 20 is low noise because of the weightings, the processed signal 22 is a low noise version of the desired signal, having the full bandwidth or resolution.
-7- ~ 9~
The degree of artifacts in the image can depend on the nature of the combiner 21. For simplicity, the combiner 21 can simply be an adder. Here the desired selected image will have the required low noise and high resolution.
However, signal 20 contains the high frequency components of undesired structures. For example, in vessel images, signal 20 contains the edges of undesired structures such as bone and various soft tissue organs and motion artifacts.
Since these are only edges, however, a trained radiologist can ignore them and concentrate on the vessels.
This system is particularly applicab~e to the previously-described hybrld sub~raction scheme. Assume summer 13 weights the four signals to eliminate everything but iodine, while also subtracting soft tissue motion artifacts.
Summer 14 can, for example, weight the signals to provide a low noise temporal subtraction, without concern about soft tissue motion artifacts. This can be accomplished by using a large positive weight for the low energy signal before iodine is administered, and a smaller positive weight for the high energy signals since it contains less of the iodine component. These weights are reversed for the measurements taken after the iodine administration to provide a temporal subtraction of everything except iodine.
In this case, in the absence of motion, when l9 and 20 are added, the desired isolated vessel signal will be formed having its full resolution. In the presence of soft tissue motion, signal l9 will be immune and continue to be vessels only, as described in the previously referenced application on hybrid subtraction. Low noise signal 20, however, will contain motion artifacts. Thus when they are added in combiner 21, the edges of these artifacts will be present.
In accordance with the present invention, the multiple measurement noise reducing system of Macovski is modified to suppress the high frequency artifacts without undue sacrifice of the SN~. Whereas the Macovski system varies the weighting of the Mh component to handle the presence of artifact edges, the present invention varies the filtering of the lo~
frequency and high frequency signal components. In the image areas where artifact edges wou~d otherwise appear using a space invariant filter, the filter applied on the signal S has a wider spatial bandwidth and therefore the complementary filter applied on the signal component M has a narrower bandwidth. Conse~uently, the composite signal sacrifices the SNR in the artifact-edge regions, but suppresses the presence of that artifact edge because the combined signal receives relatively few of its frequency components from the artifact-laden signal M. In those areas where broad regions of artifact reside, the filter applied to the signal S possesses a relatively narrow spatial bandwidth to filter the noise significantly. Thus, the combined signal 15 obtains most of its frequency components from the lower noise signal M.
Figure 3 is a functional block diagram illustrating the noise reducing system in accordance with the invention. The system comprises a plurality of spatial filter sets 30, 31, and 32, each of which comprises a low pass filter 15, high pass filter 16, and combiner 21 (shown in Figure 1) as illustrated. Thus, the selective signal, S, 17 and the low noise signal, M, 18 are applied to each of the spatial filter sets 30-32. The appropriate filter for each picture element is selected from this bank of memories according to the degree of artifact edge present. The severity of artifact edge is measured by using an image of the artifact found in signal M. Ths image, denoted D, may be a third combination of the original measurements, such as derived by subtracting S from M with the appropriate scaling. The signal ~ is applied to an edge detection operator having N threshclds where N is the number of spatial filters in Figure 3, and the output of the edge detector operator is applied to a switch 36 for selecting the spatial filter set for eacn picture element.
Figure 4 is a functional block diagram of one embodiment of the edge detector operator 34 and includes a plurality of derivative operators 40, 41, and 4~ each r2ceiving the input imagep D, in ~h~ horizontal and ver~ical direction3 to provide the two component~ of the spatial gradient vector. The magnitud~ of the gradient vector at each point 5 a~ derived by th~ point operator 44 ar~ then compared wi~h ~hresholds for each spatial filter set. Tha higher th~
threshold level exceeded ~y the edge image, th~ wider th~
spatial bandwidth of the filter applied to ths signal S ana the narrower the spatial bandwidth of the filter applied to the signaL M. Other edge map techniques can he employed such as template filt~r~ known in th~ art ta detec~ ~dge orientation and ~trength.
There has been described an improved multiple measurement noise reducing system in which the filtering of tw~ signal 15 components ar~ spatially ~aried to minimize the ~re~ence of extraneous artifacts. By de~ecting the edge~ of an artifact through use of spatial operators and then adjusting the filtering of the two composite signal~, an improved signal with minimal distortion due to artifact edge~ i~ provided.
20 While ~he invention has been described with reference to a specific embodiment, the description is illustrative of th~
invention and i~ not to be construed as limiting the invention~
Various modifications and applications may occur to those skilled in the art without aeparting from the tru~ spirit 25 and scope of the invention as defined by the appended claims.
~J~ 3~ ~
MULTIPLE MEASUREMENT NOISE RED~CING SYS~EM USING SPACE-VARIANT FILTERS
This invention relates generally tO imaging systems such as x-ray systems, and more particularly the invention relates to the processing of images derived from a plurality of measurements.
In many imaging application areas, images are constructed as a weighted sum of a plurality of measurements. A prime example are the recent new developments in x-ray imaging.
Here measurements are made at different energies and/or at different times. These measurements are then combined to provide selective images representing specific materials.
One example is that of temporal subtraction techniques using digital radiography as described in the publication ~y C. A.
Mistretta and A. B. Crummy, "Diagnosis of Cardiovascular Disease by Digital Subtraction Angiography," in Science, Vol. 214, pp. 761-65, 1981. Here measurements taken before and after the administration of iodine into blood vessels are subtracted to provide an image of vessels alone. Another example is that of energy-selective radiography as described in the publication by L. A. Lehmann, et al., "Generalized Image Combinations in Dual KVP Digital Radiography," in Medical Physics, Vol. 8, pp. 659-67, 1931. Here measure-ments made at different energies are combined to enhance or eliminate specific materials. A third example is energy-selective computerized tomography as described in the ~L~ 3~
publication by R. E. Alvarez and A. Macovski, " Energy-Selec-tive Reconstructions in X--ray Computerized Tomography," in Physics in Medicine ~ Bioloqyr Vol. 21, pp. 733-744, 1976. Here sets of measurements are made at two energy spectra and distortion-free reconstructions are made which delinea-te the average atomic number and densi-ty of each region. A fourth e~ample is hybrid subtraction described in the publication by W. R. Brody, "Hybrid Subtraction for Improved Intravenous Arteriography," in Radioloqy, Vol. 141, pp. 828-831, 1981. Here dual energy measurements are 1() made before and after the administration of iodine. Each dual energy pair is used to elimi,nate soft tissue. In this way the resultant subtracted vessel images are immune to soft tissue motion.
In each case, where a number of measurements are combined to select specific material properties, the resultant SNR (Signal-to-noise ratio~ is reduced as compared to the non-selective image.
Therefore, these various processing techniques which improve the visualization of disease processes by selecting specific materials have the disadvantage of a reduc-tion in SNR. This reduced SNR can interfere with the ability to visualize regions of disease.
A system to improve the SNR of the hybrid subtraction and other multiple measurement systems has been developed by A. Macovski. An array of measurements are taken of an object under different conditions. These are combined to select a specific aspect of the object. The same measuremen-ts are then combined in different weightings to provide a lower-noise image, wi-thout the desired selectivity. The improved selective image is formed by combining the selective image with the lower-noise image - 2a - ~ 9 using the lower frequency components of ~he former and the higher frequency components of the latter. To assure the proper amplitude of the higher frequency componen-ts, the components are weighted such as with the ratio of the derlvative of the selective image to the derivative of the lower-noise image.
Nonlinear thresh~lds are used for those cases where this ratio becomes an unreliable indicator of the correct amplitude of the high frequency components. One limitation of tne Macovski system is the introduction of extraneous signals (herein termed "artifacts"3 by the low noise signal.
The present invention is directed to an improvement in the Macovski system by minimizing the artifacts introduced in the displayed image. In accordance with the invention, a space-variant filter is applied to a selective image signal, S, and the complementary space-variant filter to a lower noise image signal, M. The sum of these two filtered images appropriately scaled results in the improved image S. This process still combines the low frequency components of S
wi~h the high frequency components of M, but the spatial frequency distribution of the filter response varies with the picture element. The procedure may be expressed by S = Sl(*) + Mh(*) where Sl(*) and Mh(*) represent the two filtered images, l(*) the space-variant filter, and h(*) the complementary space-variant filter. Because the two filters are always complementary and the combination is linear, an important property is that the desired signal component is never distorted in the processed image. In the origina~ NRS, distortion is possible in the form of blurring when the weighting of the high pass filtered M image ~Mh), which contains the edge information, becomes small to prevent artifact edges from entering the enhanced image. In the improved system, the filter used at each picture element is chosen to suppress the amount of high frequency artifact that leaks into the processed image without undue sacrifice of the SNR.
Accordingly, an object of the present invention is a multiple measurement noise reducing system which minimizes e~traneous ~:1 3~
artifacts.
Another object of -the inventi.on is a method of improving the image in a mul.tiple measurement noise reducing system.
According to one aspect of the invention there is provided for use in an imaging system, a method of reducing extraneous signals in a first processed image signal of an object representing a desired parameter derived from a plurali-ty of measurements comprising the steps of: processing said plurali-ty of measurements to provide a second processed image signal having greater signal--to-noise ratio than said first processed signal, low pass filtering said first processed image signal to reduce its noise, high pass filtering said second processed image signal, said high pass filtering being complementary to said low pass filtering; varying the filter frequencies of said low pass fiitering and said high pass filtering in response to presence of artifacts, and combining said filtered first processed image and said filtered second processed image to provide an improved processed image signal.
Also, according to the invention, there is provided in an imaging system, apparatus for reducing ex-traneous image signals in a processed image signal of an object representing a desired parameter derived from a plurality of measurements comprising:
first processing means for processing said plurality of measure-ments and producing a first processed image signal representing said desired parameter, second processing means for processing said plurality of measurements and producing a second processed image signal having a greater signal-to-noise ratio than said firs-t processed image signal, first fil-ter means for receiving - 4a -and low pass filtering said first processed image signal, second filter means for receiving and high pass filtering said second processed imaye signal, means for varying the filter band widths of said low pass filter means and said high pass filter means in response to presence of extraneous image signals, and means for combining said filtered first processed image signal and said filtered second processed image signal.
The inven-tion and objects and features thereof will be more readily apparent from the following detailed description and appended claims when taking with the drawing, in which:
Figure 1 is a block diagram of an x-ray system for acquiring signals useful in the invention.
Figure 2 is a block diagram of an x-ray system in accordance with -the prior art.
Figure 3 is a functional block diagram of an x-ray system in accordance with the present invention.
Figure 4 is a functional block diagram of edge detection circuitry useful with the system of Figure 3.
Referring now to the drawings, Figure 1 is a block diagram of an x-ray system which includes an x-ray tube 8 supplied by variable voltage power source 3. The x-rays are transmitted through object 1, normally the human anatomy. The transmitted rays are received by x-ray detector 4, such as an image intensi-fier and television camera, or a one- or two-dimensional array of scintillators, etc. The image signal is applied to storage system 6 where a plurality of images 9, 10, 11, 12, etc., can be stored.
In temporal sub-traction a first image 9 is s-tored. Follow-~ 3 ~7~ 9 ing administration of a contrast agen-t using syringe 2, a second image 10 is stored. These are -then used in the subsequent processing systems. In energy selec-tive systems, images 9, 10, 11, 12, e-tc., are stored a-t different values of beam A-38870~HKW
-5- ~ 3~5~
energy corresponding to different anode voltages 3. In dual-energy systems, two voltages are used. In addition different x-ray beam filters, not shown, can be ad~ed.
In the Brody hybrid subtraction s~stem previously described, measurements at two voltages are taken prior to the admini-stration of contrast material. These are stored as 9 and 10. Following administration of the contrast agent, two additional measurements at the same two voltages are taken and stored as 11 and 12. ThesP four measurements are then processed to obtain a low-noise image of the vessels opacified by the contrast material from syringe 2.
Referring now to ~igure 2, the lines 9, 10, 11 and 12 repre-senting an array of x-ray measurements of an object, are processed in accordance with the Macovski system described above. These signals could be obtained, for example, from the output of the television camera in a fluoroscopic system where an array of images are formed and stored, for example, on a magnetic disc or in a solid-state frame memory. These signals can represent x-ray images taken with different photon energy spectra or taken before and after the admini-stration of a contrast agent.
In general the multiple measurements are taken so as to provide the isolation or elimination of specific materials.
One simple example involves the isolation of vessel images by subtracting measurements taken before and after the administration of an iodinated contrast agent. Other examples include measurements at different x-ray photon ener~y spectra which are then combined to eliminate or enhance specific materials.
Weighted summer 13 takes the measurements and provides selective signal 17, containing the desired material pro-perties. Unfortunately, however, signal 17 often has relatively poor SNR. The weighted sum usuall~v involves negative weights which diminish the signal. The noise variances of each independent measurement add, however, to A-3887~)-HKW
-6~
provide a large ~ariance, hence a low SNR. This SNR is considerably improved using low pass filter 15 which reduces the bandwidth. This noise reduction, however, i5 accom-plished at the price of reduced resolution. A primary objective of this invention is to restore the resolution using a source of high frequency information with reduced noise.
Signals 9, 10, 11 and 12 are also applied to weighted summing structure 14. These weights can be chosen to maximize the signal-to-noise ratio for a particular component of the image. For example, for vessel imaging they can be chosen to maximize the sisnal-to-noise ratio for the iodinated contrast agent. This is in sharp distinction to weighted summing structure 13 where the weights are chosen to isolate iodine and eliminate the various tissue components. Ideally, the weights which ma~imize the SNR will vary in different portions of the image, depending on the intervening tissue.
Thus, the weights can be varied as the signals axe scanned.
However, for simplicity, the weights can be fixed to maximize the average SNR with relatively small degradation in per-formance.
The high SNR weighted-sum signal, 18, is applied to high pass filter 16 which extracts the high frequency components.
A preferred embodiment of this high pass filter is the complement of low pass filter 15. Thus the sum of the normalized transmissions equals unity within the frequencies of interest. These filters can either be implemented in the frequency domain or as convolution functions.
Combiner 21 combines the low frequency image or signal 19 with the high frequency image or slgnal 20 to obtain the processed image 22 which is displayed in display 23. Since signal 19 is low noise because of the filtering, and signal 20 is low noise because of the weightings, the processed signal 22 is a low noise version of the desired signal, having the full bandwidth or resolution.
-7- ~ 9~
The degree of artifacts in the image can depend on the nature of the combiner 21. For simplicity, the combiner 21 can simply be an adder. Here the desired selected image will have the required low noise and high resolution.
However, signal 20 contains the high frequency components of undesired structures. For example, in vessel images, signal 20 contains the edges of undesired structures such as bone and various soft tissue organs and motion artifacts.
Since these are only edges, however, a trained radiologist can ignore them and concentrate on the vessels.
This system is particularly applicab~e to the previously-described hybrld sub~raction scheme. Assume summer 13 weights the four signals to eliminate everything but iodine, while also subtracting soft tissue motion artifacts.
Summer 14 can, for example, weight the signals to provide a low noise temporal subtraction, without concern about soft tissue motion artifacts. This can be accomplished by using a large positive weight for the low energy signal before iodine is administered, and a smaller positive weight for the high energy signals since it contains less of the iodine component. These weights are reversed for the measurements taken after the iodine administration to provide a temporal subtraction of everything except iodine.
In this case, in the absence of motion, when l9 and 20 are added, the desired isolated vessel signal will be formed having its full resolution. In the presence of soft tissue motion, signal l9 will be immune and continue to be vessels only, as described in the previously referenced application on hybrid subtraction. Low noise signal 20, however, will contain motion artifacts. Thus when they are added in combiner 21, the edges of these artifacts will be present.
In accordance with the present invention, the multiple measurement noise reducing system of Macovski is modified to suppress the high frequency artifacts without undue sacrifice of the SN~. Whereas the Macovski system varies the weighting of the Mh component to handle the presence of artifact edges, the present invention varies the filtering of the lo~
frequency and high frequency signal components. In the image areas where artifact edges wou~d otherwise appear using a space invariant filter, the filter applied on the signal S has a wider spatial bandwidth and therefore the complementary filter applied on the signal component M has a narrower bandwidth. Conse~uently, the composite signal sacrifices the SNR in the artifact-edge regions, but suppresses the presence of that artifact edge because the combined signal receives relatively few of its frequency components from the artifact-laden signal M. In those areas where broad regions of artifact reside, the filter applied to the signal S possesses a relatively narrow spatial bandwidth to filter the noise significantly. Thus, the combined signal 15 obtains most of its frequency components from the lower noise signal M.
Figure 3 is a functional block diagram illustrating the noise reducing system in accordance with the invention. The system comprises a plurality of spatial filter sets 30, 31, and 32, each of which comprises a low pass filter 15, high pass filter 16, and combiner 21 (shown in Figure 1) as illustrated. Thus, the selective signal, S, 17 and the low noise signal, M, 18 are applied to each of the spatial filter sets 30-32. The appropriate filter for each picture element is selected from this bank of memories according to the degree of artifact edge present. The severity of artifact edge is measured by using an image of the artifact found in signal M. Ths image, denoted D, may be a third combination of the original measurements, such as derived by subtracting S from M with the appropriate scaling. The signal ~ is applied to an edge detection operator having N threshclds where N is the number of spatial filters in Figure 3, and the output of the edge detector operator is applied to a switch 36 for selecting the spatial filter set for eacn picture element.
Figure 4 is a functional block diagram of one embodiment of the edge detector operator 34 and includes a plurality of derivative operators 40, 41, and 4~ each r2ceiving the input imagep D, in ~h~ horizontal and ver~ical direction3 to provide the two component~ of the spatial gradient vector. The magnitud~ of the gradient vector at each point 5 a~ derived by th~ point operator 44 ar~ then compared wi~h ~hresholds for each spatial filter set. Tha higher th~
threshold level exceeded ~y the edge image, th~ wider th~
spatial bandwidth of the filter applied to ths signal S ana the narrower the spatial bandwidth of the filter applied to the signaL M. Other edge map techniques can he employed such as template filt~r~ known in th~ art ta detec~ ~dge orientation and ~trength.
There has been described an improved multiple measurement noise reducing system in which the filtering of tw~ signal 15 components ar~ spatially ~aried to minimize the ~re~ence of extraneous artifacts. By de~ecting the edge~ of an artifact through use of spatial operators and then adjusting the filtering of the two composite signal~, an improved signal with minimal distortion due to artifact edge~ i~ provided.
20 While ~he invention has been described with reference to a specific embodiment, the description is illustrative of th~
invention and i~ not to be construed as limiting the invention~
Various modifications and applications may occur to those skilled in the art without aeparting from the tru~ spirit 25 and scope of the invention as defined by the appended claims.
Claims (13)
1. For use in an imaging system, a method of reducing extraneous signals in a first processed image signal of an object representing a desired parameter derived from a plurality of measurements comprising the steps of:
processing said plurality of measurements to provide a second processed image signal having greater signal-to-noise ratio than said first processed signal, low pass filtering said first processed image signal to reduce its noise, high pass filtering said second processed image signal, said high pass filtering being complementary to said low pass filtering;
varying the filter frequencies of said low pass filtering and said high pass filtering in response to presence of artifacts, and combining said filtered first processed image and said filtered second processed image to provide an improved processed image signal.
processing said plurality of measurements to provide a second processed image signal having greater signal-to-noise ratio than said first processed signal, low pass filtering said first processed image signal to reduce its noise, high pass filtering said second processed image signal, said high pass filtering being complementary to said low pass filtering;
varying the filter frequencies of said low pass filtering and said high pass filtering in response to presence of artifacts, and combining said filtered first processed image and said filtered second processed image to provide an improved processed image signal.
2. The method as defined by Claim 1 wherein said step of varying the filter frequencies includes the step of deter-mining the spatial location of extraneous artifact edges in said second processed image signal and increasing the bandwidth frequency of said low pass filtering and reducing the bandwidth of said high pass filtering when an artifact edge is detected.
3. The method as defined by Claim 2 wherein said step of determining the spatial location of extraneous artifact edges includes determining the components of the spatial gradient signal vector of the difference between said first processed image signal and said second processed image signal.
4. The method as defined by Claim 1 wherein said first processed image signal is derived from a first weighted sum of said measurements and said second processed image signal is derived from a second weighted sum of said measurements.
5. The method as defined by Claim 4 wherein said imaging system comprises an x-ray system and said measurements are at different x-ray energies.
6. The method as defined by Claim 4 wherein said imaging system comprises an x-ray system and said measurements are at different times.
7. In an imaging system, apparatus for reducing extraneous image signals in a processed image signal of an object representing a desired parameter derived from a plurality of measurements comprising:
first processing means for processing said plurality of measurements and producing a first processed image signal representing said desired parameter, second processing means for processing said plurality of measurements and producing a second processed image signal having a greater signal-to-noise ratio than said first processed image signal, first filter means for receiving and low pass filtering said first processed image signal, second filter means for receiving and high pass filtering said second processed image signal, means for varying the filter bandwidths of said low pass filter means and said high pass filter means in response to presence of extraneous image signals, and means for combining said filtered first processed image signal and said filtered second processed image signal.
first processing means for processing said plurality of measurements and producing a first processed image signal representing said desired parameter, second processing means for processing said plurality of measurements and producing a second processed image signal having a greater signal-to-noise ratio than said first processed image signal, first filter means for receiving and low pass filtering said first processed image signal, second filter means for receiving and high pass filtering said second processed image signal, means for varying the filter bandwidths of said low pass filter means and said high pass filter means in response to presence of extraneous image signals, and means for combining said filtered first processed image signal and said filtered second processed image signal.
8. Apparatus as defined by claim 7 wherein said first filter means and said second filter means are complementary in frequency.
9. Apparatus as defined by claim 8 wherein said means for varying the filter frequencies includes means for determining the spatial location of edges of extraneous images in said second processed image signal and increasing the bandwidth frequencies of said low pass filter means and decreasing the bandwidth frequencies of said high pass filter means when an edge of an extraneous image is detected.
10. Apparatus as defined by claim 9 wherein said means for determining the spatial location of edges includes differential means for determining the components of the spatial gradient signal vector of the difference between said first processed image signal and said second processed image signal.
11. Apparatus as defined by claim 7 wherein said first processed image signal is derived from a first weighted sum of said measurements and said second processed image signal is derived from a second weighted sum of said measurements.
12. Apparatus as defined by claim 11 wherein said imaging system comprises an x-ray system and said measurements are at different x-ray energies.
13. Apparatus as defined by claim 11 wherein said imaging system comprises an x-ray system and said measurements are at different times.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US06/468,657 US4503461A (en) | 1983-02-22 | 1983-02-22 | Multiple measurement noise reducing system using space-variant filters |
| US468,657 | 1983-02-22 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA1213989A true CA1213989A (en) | 1986-11-12 |
Family
ID=23860703
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA000447768A Expired CA1213989A (en) | 1983-02-22 | 1984-02-17 | Multiple measurement noise reducing system using space-variant filters |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US4503461A (en) |
| EP (1) | EP0117155A3 (en) |
| JP (1) | JPS59229670A (en) |
| AU (1) | AU557426B2 (en) |
| CA (1) | CA1213989A (en) |
| IL (1) | IL71013A (en) |
Families Citing this family (47)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4598368A (en) * | 1982-12-27 | 1986-07-01 | Tokyo Shibaura Denki Kabushiki Kaisha | Superposed image display device |
| US4661858A (en) * | 1984-02-28 | 1987-04-28 | Canon Kabushiki Kaisha | Reading apparatus |
| JPS60222034A (en) * | 1984-04-18 | 1985-11-06 | 富士写真フイルム株式会社 | Image formation in energy subtraction |
| US4613986A (en) * | 1984-06-14 | 1986-09-23 | Ncr Canada Ltd. | Digital filter for enhancement of document images |
| US4729100A (en) * | 1984-08-28 | 1988-03-01 | Kabushiki Kaisha Toshiba | CT System which convolutes projection data with a frequency varying filter function |
| US4662379A (en) * | 1984-12-20 | 1987-05-05 | Stanford University | Coronary artery imaging system using gated tomosynthesis |
| GB2170373B (en) * | 1984-12-28 | 1989-03-15 | Canon Kk | Image processing apparatus |
| JPS61201372A (en) * | 1985-03-02 | 1986-09-06 | Toshiba Corp | Image processor |
| US4697594A (en) * | 1985-08-21 | 1987-10-06 | North American Philips Corporation | Displaying a single parameter image |
| GB2181330B (en) * | 1985-09-26 | 1990-05-09 | Toshiba Kk | X-ray inspection apparatus |
| US4827528A (en) * | 1985-11-15 | 1989-05-02 | Stanford University | Error-minimizing noise-reduction system |
| US4860373A (en) * | 1986-03-31 | 1989-08-22 | General Electric Company | Location dependent signal processor |
| US4792900A (en) * | 1986-11-26 | 1988-12-20 | Picker International, Inc. | Adaptive filter for dual energy radiographic imaging |
| US4761819A (en) * | 1987-02-27 | 1988-08-02 | Picker International, Inc. | Adaptive noise reduction filter for reconstructed images |
| US4887609A (en) * | 1987-05-13 | 1989-12-19 | The Methodist Hospital System | Apparatus and method for filtering electrocardiograph signals |
| US4907156A (en) * | 1987-06-30 | 1990-03-06 | University Of Chicago | Method and system for enhancement and detection of abnormal anatomic regions in a digital image |
| JP2536024B2 (en) * | 1988-02-29 | 1996-09-18 | 株式会社島津製作所 | X-ray image processing device |
| JP2536094B2 (en) * | 1988-10-20 | 1996-09-18 | 富士ゼロックス株式会社 | Color image processing device |
| JPH02114776A (en) * | 1988-10-25 | 1990-04-26 | Toshiba Corp | X-ray diagnostic device |
| JP2670635B2 (en) * | 1989-10-19 | 1997-10-29 | 富士写真フイルム株式会社 | Method and apparatus for displaying energy subtraction image |
| US5212637A (en) * | 1989-11-22 | 1993-05-18 | Stereometrix Corporation | Method of investigating mammograms for masses and calcifications, and apparatus for practicing such method |
| US5131049A (en) * | 1989-12-08 | 1992-07-14 | Xerox Corporation | Identification, characterization, and segmentation of halftone or stippled regions of binary images by growing a seed to a clipping mask |
| JPH0834544B2 (en) * | 1990-08-14 | 1996-03-29 | 株式会社東芝 | Image restoration processing method |
| FR2671255B1 (en) * | 1990-12-28 | 1993-03-19 | Gen Electric Cgr | METHOD AND SYSTEM FOR ELIMINATING ADVERSE SPACE FREQUENCIES IN THE RADIOLOGICAL IMAGE OF AN OBJECT. |
| US5273040A (en) * | 1991-11-14 | 1993-12-28 | Picker International, Inc. | Measurement of vetricle volumes with cardiac MRI |
| US5243444A (en) * | 1992-03-25 | 1993-09-07 | Xerox Corporation | Image processing system and method with improved reconstruction of continuous tone images from halftone images including those without a screen structure |
| FR2708166A1 (en) * | 1993-07-22 | 1995-01-27 | Philips Laboratoire Electroniq | A method of processing digitized images for the automatic detection of stenoses. |
| JP3372611B2 (en) * | 1993-10-18 | 2003-02-04 | キヤノン株式会社 | Video transmission system, video processing device, and video processing method |
| US5818517A (en) * | 1993-11-12 | 1998-10-06 | Northern Telecom Limited | Broadband interference reduction |
| US5852475A (en) * | 1995-06-06 | 1998-12-22 | Compression Labs, Inc. | Transform artifact reduction process |
| IL119283A0 (en) | 1996-09-19 | 1996-12-05 | Elscint Ltd | Adaptive filtering |
| RU2150146C1 (en) * | 1998-09-03 | 2000-05-27 | Семенченко Михаил Григорьевич | Method for image processing |
| US6233060B1 (en) | 1998-09-23 | 2001-05-15 | Seiko Epson Corporation | Reduction of moiré in screened images using hierarchical edge detection and adaptive-length averaging filters |
| US6259822B1 (en) * | 1998-10-30 | 2001-07-10 | Eastman Kodak Company | Edge enhancement which reduces the visibility of false contours |
| US6721458B1 (en) | 2000-04-14 | 2004-04-13 | Seiko Epson Corporation | Artifact reduction using adaptive nonlinear filters |
| US6633683B1 (en) * | 2000-06-26 | 2003-10-14 | Miranda Technologies Inc. | Apparatus and method for adaptively reducing noise in a noisy input image signal |
| US6438195B1 (en) * | 2001-01-26 | 2002-08-20 | Ge Medical Systems Global Technology Company, Llc | Methods and apparatus for compensating for view aliasing artifacts |
| US7418149B2 (en) * | 2004-07-30 | 2008-08-26 | Algolith Inc. | Apparatus and method for adaptive 3D noise reduction |
| GB2442256A (en) * | 2006-09-28 | 2008-04-02 | Tandberg Television Asa | Position-dependent spatial filtering |
| TWI330036B (en) * | 2006-10-27 | 2010-09-01 | Quanta Comp Inc | Apparatus for sharpening an image, and method thereof |
| US7706497B2 (en) * | 2008-03-14 | 2010-04-27 | General Electric Company | Methods and apparatus for noise estimation for multi-resolution anisotropic diffusion filtering |
| DE102009004580A1 (en) * | 2009-01-14 | 2010-07-15 | Siemens Aktiengesellschaft | Scanning and reconstruction of a CT system and CT system |
| US8798388B2 (en) * | 2009-12-03 | 2014-08-05 | Qualcomm Incorporated | Digital image combining to produce optical effects |
| WO2012037151A2 (en) * | 2010-09-13 | 2012-03-22 | University Of Southern California | Efficient mapping of tissue properties from unregistered data with low signal-to-noise ratio |
| JP6076981B2 (en) * | 2011-08-19 | 2017-02-08 | コーニンクレッカ フィリップス エヌ ヴェＫｏｎｉｎｋｌｉｊｋｅ Ｐｈｉｌｉｐｓ Ｎ．Ｖ． | Frequency dependent composite of X-ray images of different modalities |
| TWI596573B (en) | 2013-04-25 | 2017-08-21 | 財團法人工業技術研究院 | Image processing device for reducing image noise and the method thereof |
| CN111089871B (en) * | 2019-12-12 | 2022-12-09 | 中国科学院苏州生物医学工程技术研究所 | Phase information separation method and system of X-ray grating phase contrast image, storage medium and equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4445226A (en) * | 1981-05-05 | 1984-04-24 | The Board Of Trustees Of The Leland Stanford Junior University | Multiple-energy X-ray subtraction imaging system |
| NL8105223A (en) * | 1981-11-18 | 1983-06-16 | Philips Nv | DEVICE FOR DIFFERENCE IMAGE DETERMINATION. |
| US4463375A (en) * | 1982-09-07 | 1984-07-31 | The Board Of Trustees Of The Leland Standford Junior University | Multiple-measurement noise-reducing system |
-
1983
- 1983-02-22 US US06/468,657 patent/US4503461A/en not_active Expired - Fee Related
-
1984
- 1984-02-17 CA CA000447768A patent/CA1213989A/en not_active Expired
- 1984-02-20 AU AU24731/84A patent/AU557426B2/en not_active Ceased
- 1984-02-21 EP EP84301119A patent/EP0117155A3/en not_active Withdrawn
- 1984-02-21 IL IL71013A patent/IL71013A/en unknown
- 1984-02-22 JP JP59032255A patent/JPS59229670A/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| AU2473184A (en) | 1984-08-30 |
| US4503461A (en) | 1985-03-05 |
| IL71013A (en) | 1988-11-15 |
| AU557426B2 (en) | 1986-12-18 |
| IL71013A0 (en) | 1984-05-31 |
| EP0117155A2 (en) | 1984-08-29 |
| JPS59229670A (en) | 1984-12-24 |
| EP0117155A3 (en) | 1987-02-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA1213989A (en) | Multiple measurement noise reducing system using space-variant filters | |
| US4499493A (en) | Multiple measurement noise reducing system using artifact edge identification and selective signal processing | |
| EP0105619B1 (en) | Multiple-measurement noise-reducing system | |
| EP1328903B1 (en) | Method and apparatus for digital image defect correction and noise filtering | |
| US4792900A (en) | Adaptive filter for dual energy radiographic imaging | |
| JP4598507B2 (en) | System and method for image noise reduction using minimum error space-time recursive filter | |
| US6792159B1 (en) | Correction of defective pixels in a detector using temporal gradients | |
| CN101510298A (en) | Synthesis correction method for CT pseudo-shadow | |
| US6760401B2 (en) | Apparatus and method for processing of digital images | |
| KR20160089976A (en) | X-ray imaging apparatus and control method for the same | |
| US10475181B2 (en) | Image processing apparatus, image processing method, and image processing program | |
| Levin et al. | Digital subtraction angiography: principles and pitfalls of image improvement techniques | |
| US7889904B2 (en) | Image processing device, image processing method, program, storage medium and image processing system | |
| US7068854B1 (en) | Correction of defective pixels in a detector | |
| US10438327B2 (en) | Image processing apparatus, image processing method, and image processing program | |
| US6697663B1 (en) | Method and apparatus for reducing noise artifacts in a diagnostic image | |
| Macovski et al. | Measurement-dependent filtering: a novel approach to improved SNR | |
| CN105593899A (en) | Network configuration using service identifier | |
| Anas et al. | Removal of ring artifacts in x-ray micro tomography using polyphase decomposition and spline interpolation | |
| Macovski | Iodine imaging using spectral analysis | |
| WO2009065441A1 (en) | Method and arrangement in fluoroscopy and ultrasound systems | |
| Kruger | Image data acquisition, processing, storage and display | |
| Nishimura | Signal processing methods for intravenous angiography | |
| Niv et al. | Data predictability for compression of digital fluorography images | |
| Giger | INVESTIGATION OF BASIC IMAGING PROPERTIES IN DIGITAL RADIOGRAPHY. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| MKEX | Expiry |