POST-PROCESSING METHOD FOR
REDUCING ARTIFACTS IN BLOCK-CODED
DIGITAL IMAGES, AND POST-PROCESSING
DEVICE FOR ACTUATING SUCH METHOD
CROSS REFERENCE TO RELATED
This is a continuation-in-part of U.S. patent application Ser. No. 08/584,529, filed Jan. 11, 1996, which issued as U.S. Pat. No. 5,870,495 on Feb. 9, 1999.
The present invention relates digital image coding. More precisely, the invention relates to a post-processing method for reducing artifacts in block-coded digital images, and to a post-processing device suitable for actuating such a method.
BACKGROUND OF THE INVENTION
With the diffusion of digital communication systems, digital images are more and more used. This has led to the diffusion of still and video cameras with digital acquisition and processing capability.
In order to better exploit storage devices and transmission bandwidth, digital image compression standards have been developed, such as JPEG for still images, and MPEG-1 and MPEG-2 for digital television image sequences.
The above-referred compression standards provide for block-coding based on Discrete Cosine Transform (DCT). A digital image is divided into blocks of pixels, and each block is encoded independently from the others. DCT coefficients for the pixels of each block are evaluated and a quantization matrix is applied to the DCT coefficients to reduce the information to be stored or transmitted. When the image is to be displayed, it must be decoded in advance.
Due to the quantization process, these image compression methods are lossy, i.e., they cause a loss of information in the decoded image with respect to the original image. The decoded image can thus present noticeable degradation, mainly consisting of two kinds of artifacts known in the art under the names of "grid noise" and "staircase noise".
In order to reduce the image degradation, post-processing methods of processing the decoded image have been proposed which allow for attenuating grid noise and staircase noise.
SUMMARY OF THE INVENTION
In view of the state of the art described, it is an object of the present invention to provide a new post-processing method for reducing artifacts in block-coded digital images.
An embodiment of the invention is directed to a postprocessing method for reducing artifacts in block-coded digital images. The method includes:
a) dividing an input image into a plurality of image blocks;
b) for each image block, estimating global features of said image block providing information on an average content of image edges along the horizontal and vertical directions of said image block;
c) for each pixel of an image block under examination, estimating local features for said pixel providing infor
mation on the content of image edges along the horizontal and vertical directions of an image area around said pixel;
d) modifying the value of said pixel according to both said 5 global features of the image block to which said pixel
belongs and said local features of the image area around said pixel. Another embodiment of the invention is directed to a post-processing device for reducing artifacts in block-coded 10 digital images. The device includes:
first means supplied with an input image for estimating global features of an image block under examination, said global features providing information on an average content of image edges along the horizontal and 15 vertical directions of said image block;
second means supplied with said input image for estimating local features for each pixel of the image block under examination, said local features providing information on the content of image edges along the hori20 zontal and vertical directions of an image area around said pixel;
third means supplied with said global features and said local features for modifying the value of said pixel according to both said global features and said local 25 features.
Features and advantages of the present invention will be made apparent from the following detailed description of an embodiment thereof, illustrated as a non-limiting example in the annexed drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic block diagram illustrating a device that implements a method according to the present inven35 tion.
FIG. 2 shows a digital image divided into image blocks. FIG. 3 shows in detail an image block of the digital image of FIG. 2.
FIG. 4 shows an array of pixels of the image block of FIG.
FIG. 5 shows an image sub-block of the image block of FIG. 3 used for evaluating global features of the image block.
FIG. 6 shows an horizontal processing window used for 45 evaluating local features in the horizontal direction for a generic pixel of the image block.
FIG. 7 shows a vertical processing window used for evaluating local features in the vertical direction for said generic pixel.
50 FIGS. 8 and 9 shows two membership functions used to perform a fuzzy computation.
FIG. 10 is a block diagram of a device according to the present invention.
FIG. 11 shows the structure of two blocks of the device of 55 FIG. 10.
FIG. 12 is a block diagram of two other blocks of the device of FIG. 10.
DETAILED DESCRIPTION OF THE 60 INVENTION
With reference to FIG. 1, there is shown a block diagram illustrating a device that implements a post-processing method according to the present invention. An input 65 decoded compressed digital image In is supplied to a Feature Extraction block FE. Block FE provides for analyzing the image to evaluate global and local features thereof. The
global and local features, respectively GF and LF, of the image In evaluated by block FE are supplied as inputs to a Fuzzy Process block FUZZY which, according to fuzzy rules, determines parameters FA suitable for determining the kind of filtering to be performed, in accordance to the global and local features GF and FL of the image. The parameters FA calculated by block FUZZY are supplied to a Filter Composition block FC which according to said parameters FA determines the type of filtering to performed out of a set of predefined filters (block FS). Filter parameters FP determined by block FC are then supplied to a Processing block PROC, also supplied directly with the input image In, which performs the filtering of the input image In according to the filter parameters FP to provide a post-processed output image Out.