KR20060103409A - Region of interest (roi) coding method and apparatus with slice structure - Google Patents

Abstract

Disclosed are a method and apparatus for encoding and decoding an image. The image encoding method may include setting the region of interest and the background region in an input image input in units of screens, and setting quantization coefficients in each region so that the image quality of the region of interest is better than that of the background region; The macroblocks corresponding to each region set in the step (b) and (a) are divided into separate slice structures, and the information necessary to decode each slice is encoded into a slice header. (C) determining each quantization coefficient, quantizing the macroblock image data according to the quantization coefficient determined in step (c) after predicting / transforming encoding, and entropy of the quantized image data in units of macroblocks. (E) encoding the. In the ROI encoding method, an input image is separated into an ROI and a background region and encoded to increase encoding efficiency, and each of the slices is transmitted in separate slices, thereby increasing transmission efficiency, but encoding quantization coefficients of each region in a slice header between the regions. The subjective image quality is improved by changing the quantization coefficient closer to the quantization coefficient in the adjacent region so that the boundary is not exposed.

Slice, DCT

Description

Method and apparatus for region of interest coding and decoding having slice structure {Region of interest (ROI) coding method and apparatus with slice structure}

The detailed description of each drawing is provided in order to provide a thorough understanding of the drawings cited in the detailed description of the invention.

1A is a block diagram schematically illustrating an embodiment of an apparatus for encoding a region of interest image having a slice structure according to the present invention.

1B is a block diagram schematically illustrating an embodiment of an apparatus for decoding a region of interest having a slice structure according to the present invention.

FIG. 2A is a flowchart illustrating an embodiment of a region of interest image encoding method performed by the apparatus of FIG. 1A.

2B is a flowchart illustrating an embodiment of a region of interest image encoding method performed by the apparatus of FIG. 1B.

3A and 3B illustrate an embodiment of a slice header.

4A and 4B illustrate a process of determining a region of interest and a background region on a screen.

5A and 5B are diagrams illustrating a process of determining a section in which a quantization coefficient is to be changed at a boundary portion of each region.

The present invention relates to a method and apparatus for encoding and decoding an image, and more particularly, to a method and apparatus for transmitting and decoding an image obtained by separating an ROI and a background region and encoding the image.

In general, when encoding and transmitting an image, since the amount of the image to be transmitted is too large, the image is divided and transmitted in a predetermined unit. An example of such a unit is the slice structure used in H.263.

When encoding an image, encoding efficiency can be increased by separating the ROI and the background region by encoding the visual characteristics of the human. This is because the image quality of the ROI is considered more important than the background image quality when a person views the image. to be.

When separating and encoding the ROI and the background, it is efficient to bundle each region in an independent unit (slice) structure and transmit it. By grouping the two areas into separate units, each area can be transmitted using a different method, for example, a unit of interest corresponds to more FEC parity packets than the background area. UEP (unequal error protection) method, such as using it may be applied.

When encoding the ROI and the background region, the ROI is quantized using small quantization coefficients to improve the image quality, while the background region uses large quantization coefficients to increase the compression ratio. In this case, one quantization coefficient is assigned to each region and encoded. In this case, the boundary between the background region and the region of interest is prominent due to the difference in the quantization coefficient value of each region. This boundary greatly impairs the subjective quality of the reconstructed image. In order to solve this problem, the quantization coefficient between the background region and the region of interest must be gradually changed, not rapidly changed.

On the other hand, in the H.263 version 2 (ITU-T Recommendation H.263, Video Coding for Low Bit Rate Communication, Jan. 1998), a group of GOBs (group of A quantization coefficient used in the GOB is encoded in each GOB header. In Annex K, a slice image is encoded by using a slice structured mode to divide a region into groups of consecutive macroblocks (MBs) or macroblock groups of arbitrary rectangular units. Each slice can be independently decoded with a slice header. The slice header starts with the start code and starts with the start code, the position in the image of the starting macroblock, the size of the slice, and the quantization coefficient used to encode the macroblock within the slice. It has a back.

Although this method can be applied to the region of interest coding, in this case, only the quantization coefficients of the corresponding regions bounded by the unit of GOB or slice are transmitted, so that the quantization coefficients of adjacent regions are not known, and thus quantization between the two regions. If the coefficient difference is large, the boundary between the two regions becomes prominent, which greatly impairs the subjective picture quality.

In addition, the MPEG-4 Part2 standard (ISO / IEC 14496-2, Information Technology-Generic Coding of Audio-Visual Objects, Part 2: Visual (Final Draft of International Standard), Dec. 1998) Can be encoded for each object, and quantization coefficients can be used for each object. However, when one object is encoded, the quantization coefficients of the object and the adjacent object or the background object are not coded together, so when applied to the ROI coding, the quantization coefficient between the object corresponding to the ROI and the object corresponding to the background region is applied. If the difference is large, the boundary becomes prominent, which greatly impairs the subjective picture quality.

In addition, U. S. Patent No. 6,256, 423 (Intra-frame quantizer selection for video compression) is a compression method that defines the region of interest, the background region, the transition region between the two regions and determine the quantization coefficient between each region. Although the transition region between the ROI and the background region can be solved to some extent, each region has independent quantization coefficient values, so that the quantization coefficient gradually changes between the ROI and the background region. This is inefficient because it requires multiple transition regions.

The technical problem to be achieved in the present invention is to transmit the image information, divided into the unit of interest area and the background area unit, but transmitted along with the quantization coefficient information of the corresponding area and the adjacent area or the background area so that the boundary between adjacent areas is not revealed. It is to provide a method and apparatus.

Another object of the present invention is to provide a recording medium in which the region of interest encoding and decoding method is recorded by program code executable by a computer.

According to an aspect of the present invention, an image encoding method includes (a) setting a region of interest and a background region in an input image input in units of screens, and quantizing each region so that the image quality of the region of interest is better than that of the background region. Setting a coefficient; (b) dividing each region set in the step (a) into independent slice structures and encoding each slice header including both the quantization coefficients of the ROI and the background region; (c) determining a quantization coefficient of each of the macroblocks corresponding to each region set in step (a); (d) quantizing the input image for each macroblock according to quantization coefficients determined in step (c) after predictive encoding and transform encoding; And (e) encoding the entropy of the quantized image data in macroblock units.

And encoding the header of the slice corresponding to the ROI by including the quantization coefficients of the background region, and encoding the header of the slice corresponding to the ROI by including the quantization coefficient of the ROI.

The method of setting the quantization coefficients may include setting the quantization coefficients such that the image quality of the ROI is better than the image quality of the background region, and quantization of adjacent regions encoded in the slice header at the boundary between the ROI and the background region. The quantization coefficient is changed to approximate the coefficient, and the section in which the quantization coefficient changes is set to be proportional to the size of each region. The region of interest encoding method may be performed on a recording medium recorded with a program code executable on a computer.

In order to accomplish the above technical problem, the ROI encoder sets a ROI and a background region in an input image input in units of screens, and sets quantization coefficients in each region so that the image quality of the ROI is better than that of the background region. A region of interest encoding modeling unit configured to determine a quantization coefficient of each of the macroblocks corresponding to each region so that the boundary between the region of interest and the background region is not exposed; A slice header encoder to separate each region determined by the ROI encoding modeling unit into independent slice structures and to include and encode quantization coefficients of the ROI and the background region in each slice header; And an entropy encoder for predicting and transform encoding the input image according to macroblocks, and quantizing the entropy according to the quantization coefficients determined by the ROI modeling unit and encoding entropy.

The header of the slice corresponding to the ROI is encoded including the quantization coefficients of the background region, and the header of the slice corresponding to the ROI is encoded including the quantization coefficient of the ROI.

The apparatus for encoding a region of interest sets a quantization coefficient such that the image quality of the region of interest is better than that of the background region, and at the boundary between the region of interest and the background region, the quantization coefficient of the adjacent region encoded in the slice header is close to the quantization coefficient. The quantization coefficient is changed so as to be changed, and the section in which the quantization coefficient is changed is set to be proportional to the size of each region.

The region of interest decoding method for achieving the technical problem comprises the steps of: (a) entropy decoding the bit stream received for each slice; (b) determining quantization coefficients of each macroblock using quantization coefficients of the magnetic region and the adjacent region included in the slice header among the information decoded in step (a); (c) decoding the image through inverse quantization / inverse transformation using the quantization coefficient determined in step (b); And (d) configuring the image of each region decoded in the step (c) into one screen by referring to the position information of each region included in the slice header.

The setting of the quantization coefficients of each macroblock may include changing the quantization coefficients so as to approximate the quantization coefficients of adjacent regions included in the slice header decoded at the boundary between the ROI and the background region, and in other regions, Use quantization coefficients.

The section for changing the quantization coefficient is set to be proportional to the size of the region, and the ROI decoding method may be performed by a recording medium recorded with program code executable on a computer.

The ROI decoding apparatus may include an entropy decoder configured to entropy decode a bit string received for each slice; A region of interest decoding modeling unit which determines quantization coefficients of each macroblock using quantization coefficients of the magnetic region and the adjacent region included in the slice header among the information decoded by the entropy decoder; An inverse quantizer for inverse quantization using the quantization coefficients determined by the ROI modeling unit; An image reconstruction unit for inversely transforming an inverse quantized signal by the inverse quantizer and reconstructing an image by compensating a predicted signal; And an image configuring unit configured to configure an image of each region restored by the image restoring unit into one screen by referring to position information of each region included in a slice header.

The ROI decoding modeling unit changes the quantization coefficients to be close to the quantization coefficients of adjacent regions included in the slice header decoded at the boundary between the ROI and the background region, and uses the quantization coefficients of the magnetic domain in other regions.

In order to fully understand the present invention, the operational advantages of the present invention, and the objects achieved by the practice of the present invention, reference should be made to the accompanying drawings which illustrate preferred embodiments of the present invention and the contents described in the accompanying drawings.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Like reference numerals in the drawings denote like elements.

1A is a block diagram schematically illustrating an embodiment of an apparatus for encoding a region of interest according to a visual characteristic of a person, according to an exemplary embodiment of the present invention. Referring to FIG. 1A, a region of interest image encoding apparatus may include a region of interest encoding modeling unit 100, a slice header encoder 101, a spatial / temporal predictive encoder 102, a transform encoder 103, and a quantizer 104. ) And an entropy encoder 105.

1B is a block diagram schematically illustrating an embodiment of an apparatus for decoding a region of interest according to a visual characteristic of a person according to the present invention. Referring to FIG. 1B, the ROI image decoding apparatus includes an entropy decoder 106, an ROI modeling unit 107, an inverse quantization unit 108, an image reconstruction unit 109, and an image configuration unit 110. do.

2A is a flowchart illustrating an embodiment of a region of interest image encoding method performed by the apparatus of FIG. 1A, and FIG. 2B is a flowchart illustrating an embodiment of a region of interest image decoding method performed by the apparatus of FIG. 1B.

1A and 2A, the ROI encoding modeling unit 100 sets an ROI and a background region in an image (video) input in units of screens, and the image quality of the ROI is better than that of the background region. A quantization coefficient is set in each region, and the quantization coefficient of each macroblock corresponding to each region is determined so that the boundary between the two regions is not exposed.

The region of interest coding modeling unit 100 sets a region of interest and a background region in an image (video) input in units of screens (step 200), and improves the image quality of the region of interest and reduces the image quality of the background region. Set the quantization coefficient of each region so as to (step 201), and change the quantization coefficient value of the macroblock close to the boundary to be close to the value of the quantization coefficient of the adjacent region so that the boundary of each region is not revealed (step 203). This value is provided to the quantization unit 104.

Since the ROI coding modeling unit 100 models the image of the ROI to have good image quality and the background image of the ROI, the quantization coefficient of the ROI is small and the quantization coefficient of the ROI is large. The determination of this quantization coefficient is determined according to the target bit amount. In addition, the ROI encoding modeling unit 100 changes the quantization coefficient value of the macroblock near the boundary so that the boundary between the regions is not exposed. A detailed description thereof will be described with reference to FIGS. 4A to 5B.

The slice header encoder 101 encodes information necessary for decoding each slice (step 202). The necessary information is included in the slice header as shown in FIGS. 3A and 3B. Referring to FIG. 3A, the slice header encodes not only the quantization coefficients to be used in the magnetic domain but also the quantization coefficients of the adjacent regions so that the quantization coefficients of the macroblocks located near the boundary with the adjacent regions are changed to be closer to the quantization coefficients of the adjacent regions. To prevent blocking between regions.

In the case of the background region, as shown in FIG. 3B, only the quantization coefficient of the magnetic region may be included and the quantization coefficient of the other region may not be included. In this case, the process of changing the quantization coefficient of the macroblock near the boundary in consideration of the quantization coefficient of the adjacent region is not performed.

The predictive encoding unit 102 predictively encodes the input image, and the transform encoding unit 103 performs a discrete cosine transform on the predictively encoded input image (step 204). In order to simplify encoding calculation, the prediction encoder 102 and the transform encoder 103 perform encoding in units of macroblocks.

The quantization unit 104 receives image data that is predictively encoded for each macroblock through the spatial / temporal predictive encoding unit 102 and transform-coded through the transform encoding unit 103, and is determined by the ROI encoding modeling unit 100. In operation 205, image data for each macroblock is quantized according to the quantization coefficients.

On the other hand, the quantization coefficients provided from the ROI coding modeling unit 100 increase in quantization coefficients from the ROI to the background area of the screen, and thus have different losses due to quantization. That is, the loss in the area of interest that receives the attention of the person is small and the loss in the background area that receives the less eye becomes large.

The entropy encoder 105 encodes the entropy of the image data quantized by the quantizer 104 and outputs the encoded entropy as a bit string (step 206).

1B and 2B, the entropy decoding unit 106 decodes the bit string received and stored in the buffer (step 207) (step 208), and decodes the information of the slice header therefrom (step 209). ) To the ROI decoding modeling unit 107. The ROI decoding modeling unit 107 sets a section for changing the quantization coefficients according to the size of the region, and determines the quantization coefficient for each macroblock by using the quantization coefficients of the magnetic region and the adjacent region, thereby dequantizing the 108. Transfer to step 210.

The image reconstructor 109 performs inverse transformation for each macroblock and compensates the predicted information to restore the image (step 212). The image reconstructor 110 reconstructs the reconstructed image for each region of the ROI and the background region. The screens are combined according to their position (step 213).

The method of setting the section in which the quantization coefficients are to be changed at the time of decoding or the method of changing the quantization coefficients is the same as that used at the time of encoding.

As described above, the ROI and the background region divided in consideration of the visual characteristics of the human are divided into respective slices and encoded or decoded, and both quantization coefficients of the two regions are encoded in each slice header, thereby gradually moving from the ROI to the background region. By controlling the quantization coefficient to be gradually increased, inter-blocking phenomenon can be effectively eliminated while improving coding efficiency.

4 illustrates an example of distinguishing an ROI and a background region from a single image. In FIG. 4A, one image includes one ROI and one background region, and FIG. 4B illustrates one image including one ROI and two ROIs. Herein, the region of interest refers to the insides of the rectangles 300, 302, and 303 indicated by thick lines, and one or more regions of interest exist in one image. The background area refers to all parts 301 and 304 except for the area of interest.

FIG. 5 illustrates an example in which sections 400, 401, and 402 for changing quantization coefficients are set so that boundaries between adjacent regions are not exposed when respective quantization coefficients are given to the ROI and the background region set as shown in FIG. 4. to be.

The size of this section may be determined in proportion to the size of each region. In the case of FIG. 4A, since one region of interest is included in one image, the quantization coefficient of the region adjacent to the region of interest in the background region may be changed in consideration of the quantization coefficient of the region of interest.

5A shows a section for changing the quantization coefficient set in this case. However, when there are a plurality of ROIs in a single image as shown in FIG. 4B, the section for changing the quantization coefficients in the background region may not be set. FIG. 5B shows such a case where it can be seen that the section in which the quantization coefficient is changed is set only within the ROI. Even in this case, the quantization coefficient of the background region may be changed by determining a region of interest closer to the background region without setting a section as shown in FIG. 5B.

After the quantization coefficients used for the ROI and the quantization coefficients used for the background region are included in each slice header as shown in FIGS. 3A and 3B and transmitted in the section set as shown in FIG. A quantization coefficient corresponding to each region is used, and the quantization coefficient is changed to be close to the quantization coefficient of the adjacent region so that the boundary between the two regions is not exposed in the interval set in FIG. 5.

A method of setting a section to change the quantization coefficient or a method of changing the quantization coefficient within this section may additionally not encode information on the method using the same method in the encoder and the decoder.

The ROI encoding and decoding method having a slice structure according to the present invention improves the subjective image quality by increasing the image quality of the ROI but not revealing a boundary with the background. Since the ROI and the background region are separated and encoded into different slices, the slices may be transmitted using different methods to protect the image quality of the ROI in case of an error during transmission.

The invention can also be embodied as computer readable code on a computer readable recording medium. The computer-readable recording medium includes all kinds of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disks, optical data storage devices, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

The best embodiments have been disclosed in the drawings and specification above. Although specific terms have been used herein, they are used only for the purpose of describing the present invention and are not used to limit the scope of the present invention as defined in the meaning or claims. Therefore, those skilled in the art will understand that various modifications and equivalent other embodiments are possible from this. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

As described above, the method and apparatus for encoding and decoding a region of interest having a slice structure according to the present invention divides a screen into a region of interest and a background region and encodes each into separate slices, but does not rapidly change the quality difference between the two regions. There is an effect of improving image quality by encoding the quantization coefficients in the adjacent region.

In addition, according to the present invention, a method and apparatus for encoding and decoding a region of interest having a slice structure may gradually change the quantization coefficient between the background region and the region of interest when encoding the region of interest and the background region into separate units (slices). In this method, the quantization coefficients used in each region (slice) as well as the quantization coefficients to be used in other adjacent regions are encoded and transmitted together. By changing to close to the coefficient value, there is an effect of eliminating the phenomenon in which the boundary between regions becomes prominent.

Claims (20)

(a) setting a region of interest and a background region in the input image input on a screen-by-screen basis, and setting quantization coefficients of each region in which the image quality of the region of interest is better than the image quality of the background region; (b) dividing each region set in the step (a) into independent slice structures and encoding each slice header including both the quantization coefficients of the ROI and the background region; (c) determining a quantization coefficient of each of the macroblocks corresponding to each region set in step (a); (d) quantizing the input image for each macroblock according to quantization coefficients determined in step (c) after predictive encoding and transform encoding; And (e) encoding an entropy of quantized image data in units of macroblocks. The region of interest encoding method of claim 1, wherein the header of the slice corresponding to the region of interest is encoded by including a quantization coefficient of the background region. The method of claim 1, wherein the header of the slice corresponding to the background area is encoded by including a quantization coefficient of the ROI. The method of claim 1, wherein the method of setting the quantization coefficients sets the quantization coefficients such that the image quality of the ROI is better than that of the background region. 5. The method of claim 4, wherein the boundary between the region of interest and the background region changes the quantization coefficients so as to be close to the quantization coefficients of adjacent regions encoded in the slice header. The method of claim 5, wherein the section for changing the quantization coefficient is set to be proportional to the size of each region. A recording medium which records the region of interest encoding method according to any one of claims 1 to 6, in a program code executable on a computer. In the input image input in units of screens, the ROI and the background region are respectively set, and the quantization coefficient is set in each region so that the image quality of the ROI is better than the image quality of the background region, and the boundary between the ROI and the background region is not revealed. A region of interest encoding modeling unit configured to determine quantization coefficients of each of the macroblocks corresponding to each region so as to prevent the quantization coefficients; A slice header encoder to separate each region determined by the ROI encoding modeling unit into independent slice structures and to include and encode quantization coefficients of the ROI and the background region in each slice header; And an entropy encoder configured to predict and encode entropy according to the quantization coefficients determined by the ROI modeling unit after predictive encoding and transform encoding the input image for each macroblock. The apparatus of claim 8, wherein the header of the slice corresponding to the region of interest includes the quantization coefficient of the background region to be encoded. The apparatus of claim 8, wherein the header of the slice corresponding to the background region is encoded by including a quantization coefficient of the region of interest. The apparatus of claim 8, wherein the ROI encoding apparatus sets quantization coefficients such that the image quality of the ROI is better than that of the background region. 12. The apparatus of claim 11, wherein the boundary region between the region of interest and the background region changes the quantization coefficients so as to approximate the quantization coefficients of adjacent regions encoded in the slice header. The apparatus of claim 12, wherein the section in which the quantization coefficient is changed is set to be proportional to the size of each region. (a) entropy decoding the bit string received for each slice; (b) determining quantization coefficients of each macroblock using quantization coefficients of the magnetic region and the adjacent region included in the slice header among the information decoded in step (a); (c) decoding the image through inverse quantization / inverse transformation using the quantization coefficient determined in step (b); And and (d) configuring the image of each region decoded in the step (c) into one screen by referring to the position information of each region included in the slice header. 15. The method of claim 14, wherein the setting of the quantization coefficients of each macroblock comprises changing the quantization coefficients so as to approximate the quantization coefficients of adjacent regions included in the slice header decoded at the boundary between the ROI and the background region. A region of interest decoding method using quantization coefficients of a magnetic region in an outside region. The method of claim 15, wherein the section for changing the quantization coefficient is set to be proportional to the size of the region. A recording medium in which the method for decoding a region of interest according to any one of claims 14 to 16 is recorded with a program code executable on a computer. An entropy decoding unit for entropy decoding the received bit string for each slice; A region of interest decoding modeling unit which determines quantization coefficients of each macroblock by using quantization coefficients of the magnetic region and the adjacent region included in the slice header among the information decoded by the entropy decoder; An inverse quantizer for inverse quantization using the quantization coefficients determined by the ROI modeling unit; An image reconstruction unit for inversely transforming an inverse quantized signal by the inverse quantizer and reconstructing an image by compensating a predicted signal; And an image constituting unit configured to configure an image of each region reconstructed by the image reconstructor into one screen by referring to position information of each region included in a slice header. 19. The apparatus of claim 18, wherein the ROI decoding modeling unit changes the quantization coefficients so as to be close to the quantization coefficients of adjacent regions included in the slice header decoded at the boundary between the ROI and the background region. A region of interest decoding apparatus using quantization coefficients. 20. The apparatus of claim 19, wherein the section for changing the quantization coefficient is set to be proportional to the size of the region.

Images