Method of removing distortions from digital images

FIELD: information technology.

SUBSTANCE: size of coding unit relative the required printer resolution is evaluated. If the size of the unit is discernible by the human eye, the following steps are carried out: the approximate metric of discernibility of a distortion caused by the Gibbs effect is determined for each coding unit and stored in memory; the approximate metric of discernibility of block distortion is determined for each border of the unit; if the corresponding element of the approximate metric of discernibility of unit distortion exceeds a predefined threshold value, a filter which can suppress block distortions is applied to the given border of the unit; for each coding unit, if the corresponding element of the approximate metric of discernibility of distortion caused by the Gibbs effect exceeds a predefined threshold value, a filter which can suppress distortions is applied to the coding unit in order to suppress distortions caused by the Gibbs effect.

EFFECT: preservation of image sharpness.

7 cl, 6 dwg

 

The invention relates to digital photography and more specifically to methods effectively eliminate artifacts caused by the imperfection of the compression algorithm, based on the encoding with the help of block discrete cosine transform (DCT), known as JPEG encoding.

The procedure, known as JPEG (name JPEG stands the Joint Photographic Experts Group, i.e. the name of the Committee that developed this standard, see http://www.jpeg.org/) [1], is widely used to compress digital data representing the photographic, static image. In order to eliminate spatial redundancy, use the discrete cosine transform and quantization. First, the image is divided into small blocks, typically 8*8 pixels, each transform block using DCT coefficients in the frequency domain (DCT coefficients), and then the DCT coefficients quantuum.

Quantization is an operation which consists in the fact that the DCT coefficients are divided evenly by the quantization matrix, whose value corresponds to each frequency region. Due to the quantization value of the frequency components with small DCT coefficient becomes zero. In General, the energy is concentrated in the low frequency range of the image signal and, consequently, the high frequency component is removed by this operation.

So the m way some of the information about the image is lost in the quantization process, resulting, after recovery of the sample image by using the inverse dekodirovaniya transformation applied to the set proventown conversion factors for recovery of the encoded image, there are the so-called block distortion and the distortion caused by the Gibbs effect (sometimes called "gibsony distortion"). These block distortion appear in boundaries of adjacent blocks of the image unevenness (jump) brightness, contrast and/or color. Gibsoni distortions occur at sharp edges of brightness (the edges of objects) and are usually fuzzy, gray line near the edge. With the increase of the compression ratio for the manifestation of distortion also increases.

To minimize the distortion caused some loss of information offered many different algorithms and methods. However, most decisions based on the detection of edges, which is a tricky procedure, the optimal parameters of which are unknown or require processing of all blocks of the image, which causes significant blurring of the image.

Prior art include the following patent documents: U.S. patent№№6707952 [1], 6748113 [2], 6845180 [3], 7050649 [4], 7003173 [5], 6807317 [].

Different metrics of distortion caused by the Gibbs effect, have been proposed in the scientific literature, they are mainly based on the detection of changes in brightness and evaluation of fluctuations about large changes in brightness (contours).

There are a large number of solutions proposed for removing block distortion, in particular, in U.S. patents№№7003174 [13], 6738528 [14], 6320905 [15], 5796875 [16], 5850294 [17], 6600839 [18], in the article, List, P., Joch, A., Lainema, J., Bjontegaard, G., Karczewicz, M. "Adaptive Deblocking Filter," in IEEE Transactions on Circuits and Systems for Video Technology, vol.13, No. 7, pp 614 - 619 [19].

Recent developments in the field of machine learning suggests that methods based on amplification, can be quite effective when using weak classifiers. Moreover, the process gain can also be seen as the process of selection of attributes, if the weak classifier uses a single sign on each stage. In a number of technical solutions have been proposed the use of learning methods for classification of images or blocks of the image into blocks, exposed block distortion, and units not subject to the block distortion. Among these solutions, it is necessary to mention the following: Z.Wang, H.R.Sheikh, and ..Bovik, "No-reference perceptual quality assessment of JPEG compressed images," in Proc. IEEE Int. Conf. Image Proc., Sept. 2002 [20], as well as publishing international the th application no WO 9922509 [21].

Closest to the claimed invention symptoms are present in the technical solution described in U.S. patent No. 6983079 [22], where a solution of the block and gibsonia distortion of each restored by the decoder frame in unit low-speed encoding video. For each block of image pixels in the reconstructed frame using the value of the first coefficient, which is located in the upper left corner of the block (so-called DC coefficient), and the values of the DC coefficients of the surrounding eight neighboring blocks to predict the other (high-frequency) coefficients (so-called AC coefficients), which could be lost at the stage of quantization in the encoding process. The predicted AC coefficients are used to classify each of the restored block and attributing it either to the category of low-or high-level categories of blocks.

Then, based on the classification unit apply adaptive low-pass filtering. Stronger smoothing using a 7-point filter is applied to the blocks with low activity, where block distortion is most noticeable, and a weak three-point smoothing filter applied to high-activity blocks, where in addition to the block distortion may be distortion caused by the effect Killed the sa. Adaptive filtering reduces gibowski sound, and block distortion without introducing unwanted blur. In this solution, it is not required to detect the edges (contours), because high-level blocks detected on the basis of the predicted values of the AU. Low-pass filtering is accomplished by using a simple predefined filter or 1/8{1 1 1 2 1 1 1}, or 1/4 {1 2 1}. Distortion caused by the Gibbs effect, removed from the high-level blocks through a low-pass filter mentioned three-point filter. Such a simplified classification of the blocks leads to many false positives and to excessive blurring of the image. Simple low-pass filtering does not protect the edges.

Most of these methods require very high computing resources include the stage of detection of the edges, or require a reference (uncompressed) image, or based on too simplistic classification of image blocks, resulting in numerous false positives.

The problem to which the invention is directed, is to offer a more effective method of removing from a restored after compression of the digital image as a block distortion, and distortion caused by the s-Gibbs effect.

The technical result achieved by the claimed improved method for the removal of distortions in the processing of digital images that have been compressed using block-based DCT coding, which involves the following operations:

- assess the size of the coding block relative to the desired resolution of the printer;

- if the block size is visible to the human eye, perform the following steps:

for each coding block to determine and store in memory an approximate metric of distinctness distortion caused by the effect of the Gibbs;

for each boundary block determine the approximate metric of distinctness block distortion

- if the corresponding element approximate metric of distinctness block distortion exceeds a predefined threshold value, apply a filter configured to suppress a block distortion, to the boundary of the block;

for each coding block, if the corresponding element approximate metric visibility of the distortion caused by the Gibbs effect, exceeds a predefined threshold value, apply a filter configured to suppress distortion for block encoding to suppress the distortion caused by the Gibbs effect.

The inventive method is applicable for both grayscale and the views of, and for the restored color images.

The novelty of the claimed invention is confirmed by the following facts:

in the inventive solution does not require detection circuit;

- the probability of the presence of distortions caused by the Gibbs effect, is calculated in the frequency domain on the basis of a simple and effective formula;

- each block of the image subjected to the block distortion, process and analyze the boundaries between the compression blocks to determine whether they require treatment or not, using easy to use arithmetic operations;

- method unit of analysis is based on the classification by the method of binary Real AdaBoost classifier, based on increased and weighted voting Committee of the weak classifiers applied to the feature vector computed using proventown of DCT coefficients of the current and neighboring blocks and matrix quantization of the image.

this approach eliminates processing blocks that are not affected by distortions DCT coding, to reduce the computing time and to prevent the degradation of high-quality images;

For a better understanding of the claimed invention further provides a detailed description involving graphic materials.

Figure 1. Diagram of the major components of the system.

Figure 2. The main stage is removal from images coded on a block-by-block DCT, block distortion and the distortion caused by the Gibbs effect.

Figure 3. Details of the procedures for the detection and removal of block distortion at the boundaries of the block.

Figure 4. The basic functions in a discrete grid cosine transform, with zigzag index scan printed on top of each basis function, just as ordered block of 8*8 DCT coefficients after quantization.

Figure 5. Illustration of the process of removing block distortion.

Fig.6 shows the results of the distortions caused by DCT encoding: above is the original image from the left side to the right - enhanced image (enlarged image), bottom - the difference between the original and enhanced images.

Figure 1 shows the functional diagram of the main components of the device that implements the inventive method. The operation of the device is controlled by the CPU 101 that executes the program code recorded in the memory 102. In memory 102 stores the original grayscale or color photograph. The image is processed and sent to the display device 104. Information exchange is carried out via the bus 106 data.

Figure 2 shows the main stages of the detection of the distortions caused by the Gibbs effect. At step 201 determines noticeable whether the coding unit of the image for the human who ski eye when the print resolution. Let the print resolution is X dots per inch (typically 300-400 dpi), and the size of the coding block in pixels is 8 pixels, then the block width and height when printing is 8/X inches. If this size is indistinguishable to the human eye, the image is no longer treated as block distortion will also be invisible, and in removing the distortion is not necessary (step 205).

Otherwise, apply the following procedure. At step 202, for each coding unit of the image calculate the approximate metric of the distinctiveness of the distortions caused by the Gibbs effect. A typical case of occurrence of such distortion is characterized by the following rules:

- current block (subject to distortion) has high activity, in other words, he presents a sufficient number of non-zero DCT coefficients;

- at least one of its four neighbors (top, bottom, left or right) is a smooth block (with a small number of non-zero DCT coefficients);

the current block is compressed, and, basically, the cause of distortion is the quantization, this means that the number of zero DCT coefficients is also visible.

Consider the use of DCT basis functions and indexes zigzag scanning to JPEG encoding. Figure 4 index of zig-zag scanning napechatan on each bazi is Noah functions in the same way, as ordered after quantization blocks of 8*8 DCT coefficients. For each 8*8 block of the image being processed are the following values:

- the maximum index of non-zero DCT coefficient, if the block is arranged (arranged in a zigzag manner on the vector, imax;

- the number of nonzero DCT coefficients NDCT;

- these fair values for its eight neighbors (adjacent blocks of 8*8), top, bottom, left, right, left-top, left-bottom, right-top, right-bottom,, k=0...7 neighboring block.

Approximate metric (RM) the distinctiveness of the distortions caused by the effect of Gibbs, RM is calculated for the current coding unit of the image using the following formula:

where

;

;

(64-NDCT- the number of zero DCT coefficients in the current block. The higher the number, the higher the compression;

θ(G1<0) - ensures that the unit has the largest number of non-zero DCT coefficients among all its neighbors;

θ(G2<0) - ensures that the unit has the largest index of the last non-zero DCT coefficient among all its neighbors;

θ(imax≥5) is a condition that the frequency distortion caused by the Gibbs effect, quite Asoka.

At step 203 check each boundary between blocks in the presence of block distortion. If this distortion is detected at the boundary of the current block, then apply the procedure for removal of block distortion. The details of step 203 is presented in figure 3.

At step 204, the distortion caused by the Gibbs effect, remove each coding block of the image, where the approximate metric distinction distortion caused by the Gibbs effect, exceeds a predefined threshold value. The threshold is defined as the maximum value among all values approximate metric distinction distortion caused by the Gibbs effect, multiplied by a number less than one, in a preferred implementation of the invention, this value is 0.5.

In the case of distortions caused by the Gibbs effect, blur near strong brightness difference (contour) is effectively removed by applying a bilateral filter. The removal of these distortions is performed only in respect of blocks, where a rough metric of the distinctiveness of the distortions caused by the Gibbs effect, exceeds a predetermined threshold value. This block image is subjected to filtration using a bilateral filter (see C.Tomasi, R.Manduchi. Bilateral Filtering for Gray and Color Images, Proc. IEEE conf. on Computer Vision, 1998 [23]). Bilateral filter is applied separately: first, per the block image I filtered line-by-line then by column, to increase the processing speed. The filtered block of Ifimage is computed by applying two successive procedures:

where h is the core of the filtering range of brightness (intensity), w is the kernel filtering in the spatial domain:

Filtering in the spatial domain helps to avoid a gap between the blocks. The parameter σdset as 0.1 or may vary in some interval. In one of the preferred variants of the invention the parameter filtering range of brightness σralso set as 0.1. Bilateral filter helps to even small variations in brightness, manifested in the form of distortions caused by the Gibbs effect, and at the same time this operation contributes to the preservation of the true contours of the image. Limited use filter only blocks that are subject to distortion, helps to avoid excessive blurring of the image. If this block approximate metric distinction distortion caused by the Gibbs effect, is below a predetermined threshold, the block is ignored and the procedure for removal of distortion is not applied.

In another embodiment of the invention, the PA is Amer filter σ rthe range of brightness estimate using the calculated dispersion gibsona distortion. The variance gibsona distortion is calculated to determine the criterion of severity gibsona distortion. Makes no sense to calculate only the variance of the block, potentially affected gibbscam distortion, as it is likely that it includes a sharp drop in brightness. The aim is to measure the variance of the blur around the outline. With this purpose is the clustering of the intensity values of pixels in a potentially distorted block. Pixels closteridium inside such a block in several classes using the k-means algorithm. Perform clustering in 2, 3 and 4 class and choose the one in which the error clustering is the smallest. Calculate the average value among dispersions discovered classes. The variance of a set of values of intensities {x1.. xN} pixels is calculated according to the standard formula (<x> is the average):

The variance is calculated between the actual intensities of the pixels and the estimated smooth"surfaces, obtained by clustering. The result is the estimation of the blur caused gibbsville distortion, as the variance in areas of potential gibsonia distortion.

The process described in figure 2, can be performed with acetamides considerations. At step 202 should access the DCT coefficients. At step 203 as necessary DCT coefficients and the values of the decoded brightness, two adjacent coding blocks simultaneously. At step 204 requires values of the decoded brightness. This means that the procedure for removing artifacts JPEG in some variant of the invention can be performed during the decoding process of the image.

After removal of distortions caused by the Gibbs effect, the execution of the method completes.

Figure 3 presents the details of the detection and removal of blocking artifacts (step 203). At step 301 the condition check whether the number of non-zero DCT coefficients in both adjacent blocks is greater than a predefined threshold value (in the current implementation it is 30). If the condition 301 is made, this border does not require removal of block distortion because the compression in this case is not very large, and the inventive method applied to the next border. If the condition 301 are not met, then perform the following step 302 and consider two adjacent blocks of the image size of 8*8 pixels Bi,jand Bi-1,jfor the upper bound of the block image Bi,j(or Bi,jand Bi,j-1for the border the left side of the block image Bi,j) to proverki, is there any noticeable border between them:

(respectively)

At step 303 checks the condition over & Delta; bi,jsome predefined threshold value (3 for 8-bit images in a preferred implementation of the invention), and if so, this boundary is not affected by the block distortion.

If the condition 303 is executed, & Delta; bi,jgreater than some predefined threshold, then the method proceeds to step 304, which checks the condition 304 about is whether the unit is the number of nonzero DCT coefficients in both adjacent blocks. In this case, both of the adjacent block is represented by a constant value. If the condition 304 completed, this border is definitely needed in the procedure for removal of block distortion, step 307. If the condition 304 is not performed, the analysis continues.

At step 305 calculates the feature vector of the border. With this purpose compute two a priori constant matrix of size 64*64, L, and So Let pi,j(k,l) are (i,j)-th basis function, i, j, k, 1={0..7}. Then

The vectors of the characteristic features of the boundaries for the current block Bi,jandimages for the top and left edges of the block, respectively, is calculated as follows. Let bi,jis the block of the corresponding DCT coefficients of block B i,jimage. Then

In this case, the length of the feature vector is equal to 64*64. In a preferred embodiment of the invention, the feature vector is shortened to simplify classifier: take only the following pair for (k, l) and (m, n): (0,0), (0,1), (0,2), (0,3), (0,4), (0,5), (1,0), (1,1), (1,2), (1,3), (1,4), (2,0), (2,1), (2,2), (3,0), (3,1), (3,3), (4,0), (4,1), (4,4), (5,0). In this case, only 441 (i.e. 21*21) component of the feature vector is recorded in the memory.

In a preferred embodiment of the invention, the classification system consists of a classifier with two classes. For the case of selection of the set of training vectors belonging to two classes, we assume that {x1; y1}; ...; {xM;M} denote the set of training data, where xm∈RN- feature vector, and ym∈{-1, +1} is its classification index. The classification process is to use Real AdaBoost based on strengthening and weighted voting Committee of the weak classifiers (see Friedman J., T.Hastie, R.Tibshirani, 2000, "Additive logistic regression: A statistical view of boosting". The Annals of Statistics, 38(2):337-374 [24]. Function solution formed as the weighted sum of weak classifiers:

At step 306, the classifier is applied to the feature vector computed in step 305. If the F(x)< 0, the boundary does not need to remove block distortion, otherwise, the block distortion at this boundary are removed, step 307. Details are provided later.

Save definition (sharpness) is one of the important tasks in the process of removing block distortion. The main problem is to ensure that the difference between genuine differences of brightness (contours) and false changes in brightness caused by the block distortion. The original contours of the brightness fluctuations are much more pronounced. Figure 5 presents a diagram of two adjacent blocks of size 8*8 pixels. Let

,

where

After the coefficients are calculated, pixels and0and b0replace with0and d0respectively:

This means that if the difference between a0and b0is significant, such a circuit is genuine, k2a small value and the new value of c0depends on a1, a d0depends on b1and the path is stored. The larger the value accepts σrthe more intensive is the smoothing block boundaries. When σrvaries from 0.05 to 0.5. In a preferred embodiment of the invention σrfixed at the value the research Institute of 0.1.

Other aspects of the invention are evident from the following drawings and description of preferred variants of the invention. For the specialist in this field is obvious that it is permissible and other embodiments of the invention with the appropriate substitutions and additions, not beyond the inventive concept, the drawings and description should be considered illustrative, and not limiting materials.

The proposed method is suitable for implementation in software, digital cameras, scanners slide, camera phones, printers, or the like.

Links on terminology

25. Y.Freund, R Schapire. A decision-theoretic generalization of on-line learning and an application to boosting Source // Proceedings of the Second European Conference on Computational Learning Theory, p.23-37, (1995).

26. Zhuravlev. Correct algebra on sets correctly (heuristic) algorithms. Part I, Cybernetics. - 1977. No. 4. - P.5-17. Part II, Cybernetics. - 1977. No. 6. - P.21-27.

Part III, Cybernetics. - 1978. No. 2. - P.35-43.

27. Alexander Vezhnevets, Vladimir Vezhnevets. Boosting - Amplification of simple classifiers. Computer graphics and multimedia. Issue # 4(2)/2006], http://cgm.computergraphics.ru/content/view/112.

1. Method of removing distortions in the processing of digital images, compressed chunked transfer-coding based on the discrete cosine transform (DCT), which is reductive the following operations:
estimate the size of the coding block relative to the required printer resolution;
if the block size is visible to the human eye, perform the following steps:
for each coding block to determine and store in memory an approximate metric of distinctness distortion caused by the effect of the Gibbs;
for each boundary block determine the approximate metric of distinctness block distortion
if the corresponding element approximate metric of distinctness block distortion exceeds a predefined threshold value, apply a filter configured to suppress a block distortion, to the border of the block,
for each coding block, if the corresponding element approximate metric distinction distortion caused by the Gibbs effect, exceeds a predefined threshold value, apply a filter configured to suppress distortion for block encoding to suppress the distortion caused by the Gibbs effect.

2. The method according to claim 1, characterized in that calculate the approximate metric of the distinctiveness of the distortions caused by the Gibbs effect, for the current coding block images, based on the number of non-zero DCT coefficients, the largest index of non-zero DCT coefficient in zigzag scanning codiovan the th block of the current image and the eight coding blocks images adjacent coding blocks of the current image.

3. The method according to claim 1, characterized in that approximate metric of distinctness block distortion is calculated first using the sum of absolute values of difference of the intensities of pairs of pixels or the number of nonzero DCT coefficients in the block, and, if, when this value exceeds a predefined threshold value, apply the classifier boundaries between blocks, clarifying thus the metric of the distinctiveness of block distortion.

4. The method according to claim 3, characterized in that the classifier boundaries between blocks based on the use of binary classifier Real AdaBoost based on strengthening and weighted voting Committee of the weak classifiers.

5. The method according to claim 1, characterized in that the degree of improvement of the boundaries set depending on a sum of absolute values of a difference of intensities of pairs of pixels, between which passes the edge of the block.

6. The method according to claim 1, characterized in that the predefined threshold value for the approximate metric of the distinctiveness of the distortions caused by the Gibbs effect, calculated using the maximum value approximate metric of the distinctiveness of the distortions caused by the Gibbs effect, among all the values.

7. The method according to claim 1, characterized in that the suppression of the distortion caused by the Gibbs effect, done the Ute by applying a bilateral filter to only adjust the unit.



 

Same patents:

FIELD: physics.

SUBSTANCE: invention proposes to use an imaging model with separation of effects associated with reflecting power of the surface R and effects associated with scene illumination characteristics L, for which: quality of a recorded image is evaluated and if there is need to correct the image, noise is filtered off; a smaller copy of the image is formed; borders for subsequent contrast enhancement at the correction step are defined on the smaller copy; the luminance channel of the initial image is selected and filtered; the image is corrected in accordance with an empirical equation of the LR imaging model:

where A is the lower boundary of contrast enhancement of the smaller copy of the image; Φ and ψ are lower and upper boundaries of contrast enhancement of the converted image; JB is the brightness component of the initial image after bilateral filtering; γ is a non-linear conversion parameter and JF is the brightness component of the enhanced image; and the resultant image is converted to RGB colour space.

EFFECT: higher image quality.

7 cl, 19 dwg, 2 tbl

FIELD: information technology.

SUBSTANCE: coding device has definition apparatus for determining image area data meant for processing in order to counter reconstruction implied by granular noise arising in image data coded based on said image data and apparatus for countering reconstruction, designed for processing in order to counter reconstruction for image area data, defined using definition apparatus when coding image data, where when the said image data are coded in data unit data modules, the said definition apparatus determines unit data which form the said image data as the said image area data, and apparatus for countering reconstruction forcibly sets the orthogonal transformation coefficient to zero, which becomes equal to zero when quantisation is carried out using the said unit data, among orthogonal transformation coefficients of unit data defined using the said definition apparatus.

EFFECT: improved quality of the decoded image.

14 cl, 18 dwg

FIELD: information technologies.

SUBSTANCE: low-pass filter is used to filter interpolated video data where extent of filtering by low-pass filter is determined on the basis of boundary strength value determined for interpolated video data and adjacent video data (interpolated and/or not interpolated). Boundary strength is determined on the basis of control video data proximity for interpolated video data and adjacent video data.

EFFECT: noise reduction for interpolated data and reduction of interpolation distortion.

45 cl, 12 dwg

FIELD: information technologies.

SUBSTANCE: invention is related to the field of processing and coding of images, in particular to method for space filtration in processing and coding of images. Technical result is achieved by the fact that method of spatial filtration for coding and processing of image includes stage of definition of efficient value of blockiness to detect efficient value of code block blockiness in compliance with number of similar pixels of edge area of code block subject to filtration; stage of image real edge detection to decide whether edge of code block is actual edge of image in compliance with difference of pixel values at two sides of code blocks edge; stage of filtration for filtration of code block in mode of filtration, which corresponds to efficient value of code block blockiness, when edge of code block is not an actual edge of image.

EFFECT: adaptive removal of blockiness artifacts in even area of image.

10 cl, 4 dwg

FIELD: physics; image processing.

SUBSTANCE: invention relates to video compression systems, particularly to a filter for eliminating modularity. A video encoder is proposed, which is based on several layers, which uses the filter for eliminating modularity, containing: apparatus for encoding the input frame; apparatus for decoding the said encoded frame; apparatus for selecting filtration intensity for eliminating modularity in accordance with whether a unit, included in the decoded frame, was coded using internal BL mode, where if, at least the current unit and the neighbouring unit were encoded using a mode other than internal BL mode, the given filtration intensity is selected with respect to the boundary between the current unit and the neighbouring unit; and if the current unit and the neighbouring unit were encoded using internal BL mode, filtration intensity, which is lower than filtration intensity, selected with respect to the boundary, is selected; and filtering apparatus for eliminating modularity with respect to the boundary between a unit and a neighbouring unit in accordance with the selected filtration intensity for eliminating modularity.

EFFECT: more accurate selection of filtration intensity for eliminating modularity taking into account internal base layer (BL) mode.

7 cl, 19 dwg

FIELD: physics; image processing.

SUBSTANCE: invention relates to processing digital images, particularly, for reducing distortions. A method is proposed for enhancing digital images, involving analysis of image parametres with subsequent extraction of brightness components, construction of a correcting filter and image correction. Images with distortions, superimposed with previous corrections, are corrected without use and analysis of parametres of the original image. Construction of the correcting filter is done on an averaged projection profile with construction of a set of local projections of brightness of the image in the direction defined by the gradient on the selected sections. The image is corrected by selecting pixels of the image in the local region, neighbouring the edges, except pixels directly on the edges. The selected sections are corrected using the constructed correcting filter with regulation of weight coefficients of the filter on local values of the gradient amplitude.

EFFECT: efficient reduction and removal of distortions in digital images with superimposed distortions.

6 dwg

FIELD: information technology.

SUBSTANCE: invention relates to digital photography, more specifically to analysis of digital image quality. The method of detecting distortions caused by Gibbs effect in JPEG coding involves evaluation of the size of the coding unit with relative the given resolution of the printing device; determination for each coding unit, whether the size of the unit makes it distinctive to the human eye with the required printing resolution, of the approximate metric of distortion distinctiveness, caused by Gibbs effect; setting to zero corresponding elements of the approximate metric of distortion distinctiveness, caused by Gibbs effect if their values are below the preferred threshold; calculation for zero elements of the approximate metric of distortion distinctiveness, caused by Gibbs effect, of the corresponding distortion dispersion; dispersion is zeroed for the rest of the elements.

EFFECT: invention can be used in detecting distortions during JPEG coding.

4 cl, 6 dwg

FIELD: physics; image processing.

SUBSTANCE: invention concerns systems of video compression, and, in particular, to the modularity elimination filter, used in the multilayered video decoder. The video decoder which is based on set of layers, using filtration for modularity elimination is offered, and the video decoder contains: agent of restoration of a video frame from an incoming bit stream; agents of choosing modularity elimination filtration intensity according to, whether there was a block included in a reconstructed frame, one is coded by means of an internal BL mode, and if, at least, of: the current block and the next block - is coded by means of the mode, differing from the internal BL mode, selection of filtration intensity in relation to border between the current block and the next block takes place; and if the current block together with the next block has been coded by means of the internal BL mode, selection of filtration intensity, which is lower than filtration intensity chosen in relation to border takes place; and an agent of modularity elimination filtration in relation to the border between the block and the next block according to the chosen modularity elimination filtration intensity.

EFFECT: maintenance of more exact choice of modularity elimination filtration intensity at the account of an internal base layer (BL) mode.

7 cl, 19 dwg

FIELD: information technology.

SUBSTANCE: present invention relates to methods and devices for processing video signals and images, and particularly to methods and devices for suppressing noise in video signals and images. The method involves selecting the first region around the current pixel; selecting two lines (horizontal and vertical) through the current pixel and pixels in the region of the current pixel; determining the horizontal gradient; determining the vertical gradient; determining the correlation matrix for the horizontal and vertical gradients, as well as eigenvectors and eigen values of the correlation matrix; determining the number of contours; determining direction of the object contour into the image; calculating weighted coefficients of the time filter; calculating adaptation coefficients of the time and bilateral filters to the power of contours, carrying out time filtering; determining the second region; dividing the second region into two sub-regions along the direction of the contour; carrying out bilateral filtration for the first and second sub-regions; calculating the result of bilateral filtration as a weighted sum of results of filtering the first and second sub-ranges; carrying out approximation of the current pixel using pixels located along the direction of the contour; calculating the final result of filtration as a weighted sum of results of bilateral filtration and approximation along the contour.

EFFECT: wider functionality and improved quality of suppressing different types of noises.

5 cl, 2 dwg

FIELD: physics, computation technology.

SUBSTANCE: invention concerns technology of video compression, particularly deblocking filters. Invention claims deblocking filter applied in videocoder/videodecoder based on multiple layers. Process of deblocking filter power (filtration power) selection during deblocking filtration in respect of margin between current block encoded in intra-BL mode and adjoining block involves determination of whether current or adjoining block has coefficients. Filter power is selected as first filter power if current or adjoining block features coefficients; and filter power is selected as second filter power if current or adjoining block does not have coefficients, So that first filter power exceeds second filter power.

EFFECT: enhanced efficiency of video deblocking.

22 cl, 13 dwg

FIELD: digital processing of images, possible use for global and local correction of brightness of digital photographs.

SUBSTANCE: system and method for correcting dark tones in digital photographs contain global contrasting module, module for conversion from RGB color system, module for determining dark tone amplification coefficient, bilateral filtration module, dark tone correction module, module for conversion to RGB color system, random-access memory block, displaying device. Global contrasting module is made with possible correction of global image contrast, module for conversion from RGB color system is made with possible conversion of image from RGB color system to three-component color system, one component of which is image brightness, and two others encode color, module for conversion to RGB color system is made with possible conversion from three-component color system, one of components of which is image brightness, and two others encode color, back to RGB color system, module for determining dark tone amplification coefficient is made with possible computation of global image brightness bar graph and can determine dark tone amplification coefficient based on analysis of signs, calculated from global image brightness bar graph, bilateral filtration module is made with possible execution of bilateral filtration of image brightness channel, dark tone correction module is made with possible correction of dark tones in image brightness channel.

EFFECT: absence of halo-effect.

2 cl, 17 dwg

FIELD: methods for removing noise in an image, possible use for improving quality of image.

SUBSTANCE: in accordance to the invention, effect is achieved due to conversion of brightness of image pixels with noise by means of solving the diffusion equation in non-divergent form, which ensures simultaneous suppression of noise and preservation of image edges.

EFFECT: simplified noise removal and increased quality of resulting digital image.

4 cl, 1 dwg

FIELD: photographic equipment engineering, image processing methods, in particular, methods for automatically correcting red-eye effect.

SUBSTANCE: method includes analysis of additional information about an image; creation of at least one array for storing image point marks; for each image point a color mark is recorded into mark array, if the color of point is a typical color for red eyes and is not a typical color for human skin; filtration of one-component image; for each point of image a boundary mark is recorded into array of marks, and also filter number is recorded into array of marks; on basis of array of marks, connected areas of points are determined; for each connected area of points with consideration of neighborhood, computation of fixed row of features; on basis of features, classification of connected areas of points onto red-eye areas and false areas; connection to red-eye connected area of image points neighboring with points of given area and close in color; connection to connected area of red-eye of image points positioned inside external contour of given area; and change of color of points in connected red-eye area.

EFFECT: ensured high quality of automatic correction of red-eye effect.

10 cl, 8 dwg

FIELD: video technology.

SUBSTANCE: invention refers to video compression, particularly to the image block compression systems. The method of image processing is suggested, which includes the definition of whether the two image blocks are adjacent or not, and whether the two blocks are subdivided or not. If the two blocks are adjacent, then the filtration of block smoothing at one or more border pixels of the two adjacent blocks occurs if it is defined that both of the adjacent blocks are not subdivided.

EFFECT: increase in the block smoothing efficiency with the use of border information.

33 cl, 24 dwg

FIELD: electric engineering.

SUBSTANCE: method includes change of image size in accordance with photocopy size and resolution of printing device, histograms are calculated for absolute values of border images, where border images are produced as a result of high-frequency filtration with convolution kernels of different size, border histogram, logarithm integral is calculated for every histogram, criteria are calculated from array of border histogram logarithm integrals, decision is taken on the basis of criteria about photograph sharpness, user is warned about possibility of printing out of focus picture, if picture as classified as out of focus.

EFFECT: detection of low quality, out of focus digital images, and their automatic exclusion from the process of printing with account of preset size of print and resolution of print.

6 cl, 4 dwg

FIELD: information technologies.

SUBSTANCE: invention refers to video compression technology, specifically to blocking effect correction filter applied in multilayered video coder/decoder. Decision mode of blocking effect correction filtration intensity is offered for frame containing block set, including the stages as follows: making decision on current block and adjacent block corrected for blocking effect; estimating whether current block and adjacent block are coded by means of internal BL mode; and if at least one either current block or adjacent block is coded by means of mode other than internal BL mode, making decision on preset filtration intensity relative to border between current block and adjacent block; and if both current block and adjacent block are coded by means of internal BL mode, making decision on filtration intensity which is lower than that chosen relative to border.

EFFECT: development of decision mode blocking effect correction filtration intensity for frame containing block set which provides proper choice of blocking effect correction filter intensity according to that whether certain block to which blocking effect correction filter is applied, uses internal base layer (BL) mode in video coder/ decoder based on layer set.

13 cl, 20 dwg

FIELD: physics; image processing.

SUBSTANCE: present invention pertains to image processing, and in particular, to the method of complexing digital multispectral half-tone images. Method of complexing digital multispectral half-tone images, including obtaining the initial images, involves breaking down each initial image to low frequency and high frequency components, separate processing of low and high frequency component images, complexing of the components, based on the principle of weighted summation for each pixel, and formation of the resultant image. Each initial image is subjected to multiple-level decomposition by the Haar wavelet through fast discrete static two-dimensional wavelet-transformation with the objective obtaining an approximate component, which is a low frequency image component, and a family of detail components, which are high frequency image components. The values of the matrix of energy characteristics of pixels are determined at all decomposition levels for each image. All detail components are filtered and the detail components are corrected through adaptive change of the values of the detail components in accordance with the inter-level dynamics of their energy characteristics. The noise microstructure is removed through adaptive threshold cut of the values of detail components on each decomposition level. The correcting brightness function and the correcting contrast function are calculated for each decomposition level, the parameter of which is a value of the approximate component. Brightness of ranges of each decomposition level is smoothed out through transformation of the approximate components by correcting brightness functions. The detail components of the contrast correcting function are transformed. A weight function is calculated for each decomposition level, the parameter of which is a value of the energy characteristic. The component of each synthesised image for each pixel at each decomposition level is calculated by weighted summation of the corresponding components of decomposing initial images using weight functions. All detail components of the synthesised image are filtered, and the detail components are corrected through adaptive change of the values of detail components in accordance with the inter-level dynamics of their energy characteristics. Noise microstructures are eliminated through adaptive threshold cut of the values of detail components at each decomposition level. The brightness correcting function and the contrast correcting function are calculated, the parameter of which is the value of approximate components of the synthesised image. The approximate component of the correcting brightness function is transformed. The detail components of the contrast correcting function are transformed. The synthesised image is formed through reconstruction using reverse fast discrete static two-dimensional wavelet-transformation, applied to the detail components of the synthesised image and approximate component of the synthesised image. The brightness range of the resulting image is matched with parameters of the video system.

EFFECT: obtaining a high quality image, containing informative image elements of the same scene, obtained in different spectral ranges.

9 dwg

FIELD: physics; processing of images.

SUBSTANCE: invention is related to the field of digital X-ray images processing. Input image is exposed to gamma-correction: extraction of square root for approximation of Poisson noise by model of additive noise distributed according to normal law; for multiplicative model of noise logarithmic conversion is performed; single-level wavelet transform of input image is done, on the basis of which wavelet coefficients are partitioned block-by-block, and standard deviation of noise for every block is assessed; prepared block ratings of noise are smoothened and interpolated by size of initial image, which gives continuously changing and locally adapting assessment of noise for the whole image; initial image is exposed to packet stationary wavelet transform by preset number of decay levels; on the basis of noise level assessment calculated at stage 2, coefficients of transform are exposed to processing with adaptive non-linear operator, which performs threshold suppression of noise and separation of image parts; reverse stationary wavelet transform is done, at that produced image with reduced level of noise and highlighted parts is exposed to reverse gamma-transform.

EFFECT: simultaneous suppression of noise and higher contrast of X-ray images.

5 dwg

FIELD: physics, processing of images.

SUBSTANCE: invention concerns numeral photo, and in particular, the analysis of quality of the numeral image. Method of revealing of unitised contortions is offered at JPEG-coding, at which: estimate the size of the coding block concerning the demanded resolution of a press; spot for each boundary of the block the approximate metric of discernability of contortion at the coding block transformation in case, the size of the coding block is a distinguishable human eye; Classify, in a case when discernability of contortions at coding transformation exceeds the predetermined threshold, boundary of the block or as boundary which demands correction for elimination of unitised contortions or as boundary which is not subject to unitised contortions, by application of the binary qualifier to the vector of the characteristic signs calculated by means of use proquantised DCT of coefficients of the adjacent blocks and a matrix of quantisation of the image.

EFFECT: increase of reliability of detection of unitised contortions at use of the underload computing and temporary resources.

4 cl, 4 dwg

FIELD: physics, computation technology.

SUBSTANCE: invention concerns technology of video compression, particularly deblocking filters. Invention claims deblocking filter applied in videocoder/videodecoder based on multiple layers. Process of deblocking filter power (filtration power) selection during deblocking filtration in respect of margin between current block encoded in intra-BL mode and adjoining block involves determination of whether current or adjoining block has coefficients. Filter power is selected as first filter power if current or adjoining block features coefficients; and filter power is selected as second filter power if current or adjoining block does not have coefficients, So that first filter power exceeds second filter power.

EFFECT: enhanced efficiency of video deblocking.

22 cl, 13 dwg

Up!