KR100483673B1 - Exceptional Sample Coding Method for Stretchable (Scalable) Coding of Binary Masks - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

Image coding method

2. The technical problem to be solved by the invention

In the conventional flexible encoding of binary masks, a large amount of bits are always allocated during ESD encoding in the process of encoding a base layer to a higher layer, so that coding efficiency is deteriorated.

3. Summary of Solution to Invention

In the flexible encoding of a binary mask that encodes a higher layer by using a base layer, when an ESD occurs during the enhancement layer encoding process, a previously generated Context-Run is generated in a range where the current Context_Run may occur. The number of bits allocated for encoding the ESD is reduced by subtracting.

4. Important uses of the invention

Applied to exceptional sample coding in binary coding of binary masks.

Description

Exceptional Sample Coding Method for Stretchable (Scalable) Coding of Binary Masks

According to the present invention, when an exceptional sample data (ESD) is generated in the scan interleaving method, a binary (scalable) encoding of a binary mask is made to reduce the bit allocation amount for representing Context_Run to increase the enhancement layer coding efficiency. It is to provide an exceptional sample encoding method.

The present invention relates to scalability coding (flexible coding) among functions supported by MPEG-4, which is currently being standardized.

In general, the next generation of video and audio coding technology and system configuration, which is developed and adopted by the MPEG Group, which has developed and decided on international standards (MPEG-1 and MPEG-2) on video and audio coding technology and system construction, International Standard (MPEG-4) is being researched and developed. The development of MPEG-4 began with the need to support the next generation of video and audio applications that are not supported by existing known standards. MPEG-4 provides new methods for communicating and accessing and manipulating video and audio data, such as object-oriented interactive functionality and access over networks of differing characteristics.

In addition, it provides characteristics that can be useful in a communication environment where errors are easily generated and a low transmission rate communication environment. Moreover, the integration of computer graphics technology provides the ability to encode and manipulate natural and artificial images and sounds together. In summary, MPEG-4 must support all the features required and expected in many applications. As a result, the expansion and openness of the multimedia information to support the functions required for all low-cost, high-performance applications possible to be newly developed or developed. Among them are transmission and storage functions and improved compression efficiency required for cost reduction. Applications currently expected to apply MPEG-4 technology include Internet Multimedia (IMM), Interactive Video Games (IVG), video conferencing and video telephony. Networked Database Service (NDB) using Interpersonal Communications, Interactive Storage Media (ISM), Multimedia Mailing (MMM), Wireless Multimedia (WMM), ATM Networks , Remote Emergency Systems (RES) and Remote Video Surveillance (RVS).

To support existing or future applications, we need image coding technology that enables users to communicate only with the objects they want, access to find and read, and edit to cut and paste. . MPEG-4, a new video and audio coding technology currently under global standardization, is designed to meet this need.

FIG. 1 is a block diagram of a MPEG-4 VOP video encoder first determined by the present International Standards Organization. This is different from the video encoder structure of H.261, H.263, MPEG-1, MPEG-2, which is the world standardized video encoding. In particular, the introduction of the concept of Shape Coder and VOP (Video Object Planes) shows the most significant difference. VOP refers to an object at one point on the time axis of arbitrary shape content that can be accessed and edited by the user, and must be encoded by VOP to support content-based functionality.

When the VOP coder inputs a VOP for each object image formed in the VOP forming unit 20 to the MOTION ESTIMATION 31, the motion estimating unit 31 performs a macroblock unit from the applied VOP. The motion is estimated.

In addition, the motion information estimated by the motion estimator 31 is input to a motion compensation unit 32 to compensate for the motion. The VOP whose motion is compensated by the motion compensator 32 is input to the subtractor 33 together with the VOP formed by the VOP forming unit 11 to detect a difference value, and the difference value detected by the subtractor 33 is The internal information of the object is encoded by the object internal encoding unit 34 in units of subblocks of the macroblock.

For example, after the X and Y axes of the macroblock are subdivided into 8 × 8 subblocks having 8 pixels each with M / 2 × N / 2, the object internal information is encoded.

On the other hand, the VOP whose motion is compensated by the motion compensator 32 and the internal information of the object encoded by the object internal encoder 34 are input to the adder 35 and added, and the output signal of the adder 35 is transferred. The previous VOP, which is input to the VEV detection unit 36, is a VOP of the image immediately before the current image is detected.

In addition, the previous VOP detected by the previous VOP detector 36 is input to the motion estimation unit 31 and the motion compensation unit 32 and used for motion estimation and motion compensation.

The VOP formed by the VOP forming unit 11 is input to a shape coding unit 37 to encode shape information.

Here, the output signal of the shape information encoder 37 varies depending on the field to which the VOP encoder is applied, and as shown by the dotted line, the output signal of the shape information encoder 37 moves the output signal of the shape information encoder 37. ), The motion compensation unit 32 and the object internal encoder 34 may be used to encode motion estimation, motion compensation, and internal information of the object.

In addition, motion information estimated by the motion estimation unit 31, object internal information encoded by the object internal encoding unit 34, and shape information encoded by the shape information encoding unit 37 are applied to the multiplexer 38. After being multiplexed and multiplexed, the bitstream is transmitted to a multiplexer which multiplexes and outputs the outputs of the plurality of encoders, although not shown in the figure.

Among the functions supported by MPEG-4 of this concept, scalability encoding is a function of transmitting and decoding a plurality of layers (base layer; enhancement layer) having different resolutions. Here, scalability refers to the capability of decoding two or more kinds of images having different spatial resolution, temporal resolution, and SNR (SNR for image quality) in one bit string.

Such scalability coding is classified into three types. They are Spatial Scalability, Temporal Scalability, SNR Scalability.

As is well known, when performing a scalability function, which is a function of transmitting and decoding a plurality of layers (base layer and high layer) having different resolutions, a lot of information is transmitted to transmit information having a plurality of different resolutions, respectively. Should be. Therefore, in order to reduce the amount of information to be transmitted, the method of estimating the high-resolution enhancement layer using the low-resolution base layer as shown in FIG.

This spatial scalability refers to a layer having a low spatial resolution as a "base layer" and a high layer as an "enhancement layer", where the base layer is encoded by a conventional MPEG-2 encoding method.

On the other hand, in the extended layer, the image of the base layer is upsampled (a method of making a high resolution screen through interpolation from the pixels of the low resolution screen) to produce an image having the same size as the high layer, and the interpolated image as well as from the high layer image. Prediction also results in more efficient encoding.

As a method of encoding a higher layer using the base layer, there are cases of using I-VOP (intra-frame coded VOP image) of the base layer and P-VOP (inter-frame forward predictive coded VOP image).

As shown in FIG. 3, two neighboring pixel values below and above are used to encode a higher layer. If two neighboring pixel values are the same, the pixel values of the current position are likely to have the same value. If the two neighboring pixel values are the same and the pixel value at the current position is the same as the two neighboring pixel values, there is no need to perform encoding.

On the other hand, when two neighboring pixel values are different, the pixel value of the current position should be encoded. This case is called a transitional sample. In addition, encoding should be performed when two neighboring pixel values are the same but the pixel values of the current position are different. This case is called an exceptional sample.

Therefore, two types of data (TSD: Transitional Sample Data, ESD: Exceptional Sample Data) exist to encode an enhancement layer. For example, when the resolution of the base layer is XxY and the resolution of the enhancement layer is (2X) x (2Y), coding in the horizontal and vertical directions should be performed. At this time, when encoding TSD, context-based Arithmetic Encoding (CAE) is performed using seven contexts in the horizontal and vertical directions, respectively.

On the other hand, in the above ESD encoding method, the ESD is encoded in the order of Context_ID, Context_Run, and ED_Length as shown in FIG. 4. Here, Context_ID refers to the context information of the location where the ESD occurred. And Context_Run refers to the number of pixels having the same Context_ID between ESDs before ESD occurs at the current position. ED_Length is the number of pixels in which ESD is continuously generated.

However, the ESD encoding method in the process of encoding the above-described conventional base layer into a higher layer has a problem in that coding efficiency is lowered because a large amount of bits are always allocated regardless of the occurrence frequency of ESD.

Accordingly, the present invention has been proposed to solve the above-mentioned problems of the prior art, and the present invention provides a higher layer by reducing the bit allocation for representing Context_Run when the exceptional sample data (ESD) occurs in the scan interleaving method. It is an object of the present invention to provide an exceptional sample encoding method in a scalable (scalable) encoding of a binary mask to improve encoding efficiency.

Hereinafter, with reference to the accompanying drawings, preferred embodiments of the present invention will be described in detail.

The video encoding system to which the present invention is applied is the same as the configuration of the MPEG-4 VOP video encoder first determined by the present International Standards Organization shown in FIG. The exceptional sample coding method in the scalable (scalable) coding method will be described with reference to FIGS. 5 to 6 as follows.

First, exceptional sample encoding (Exceptional Sampling Encoding) is composed of Context_ID, Context_Run, and ED_Length. To reduce the number of allocation bits needed to encode a Context_Run, by subtracting the previously generated Context_Run from the range in which the current Context_Run can occur, the range in which the Context_Run can occur is reduced and the bit allocation is reduced. An example is given in FIG. 5. MxN Converts ESD generation position in 2D plane into 1D. For example, assuming that it consists of an 8x8 two-dimensional plane, when converted to one-dimensional, the total length is 64. In this case, since the existence range of Context_Run for each ESD is 0 to 63, 6 bits are required for encoding. Therefore, 18 bits are required for all three ESDs.

As shown in (b) of FIG. 5, since E1 is encoded, the Context_Run for E1 is known and the range in which E2 exists can be reduced by the length thereof. Therefore, the range in which E2 may be present is 0 to 30, so that encoding may be performed using 5 bits. Since E2 is encoded using 5 bits, the decoder may also know using E1 information. Therefore, it is not necessary to send additional information. Likewise, in the case of encoding E3 as shown in FIG. 5C, since the range of E3 decreases from 0 to 10, it can be encoded using 4 bits. Thus, 6 + 5 + 4 = 15 bits are used overall to increase efficiency.

Other methods based on the above methods are also included in the present invention.

(1) Perform the above method with the one-dimensional information generated above turned upside down. In this case, one of the two methods is applied. One bit of additional information on which of the two methods is used is needed.

(2) Before applying the above method, find the Context_Run value of all ESDs and the largest value among them. Obtain the minimum bit allocation amount (MIN_BIT) to represent the value. The above method is applied, but when allocating bits to each ESD, a small amount of bits are allocated by comparing with MIN_BIT. In this case, additional information about MIN_BIT should be transmitted together. This method is advantageous when many ESDs occur at similar intervals.

(3) comparing the method of mixing the method of (2) with the method of (1), that is, the case where the one-dimensional information is used as it is when applying the method of (2) above and the case of inverting the left and right, Applies how fewer bits are generated. In addition, MIN_BIT used in the method of (2) and one bit of additional information used in the method of (1) are required.

(4) A method of moving the one-dimensional information so that the maximum length comes first and using the method of (2) described above. In other words, when you move, you think that the beginning and the end are connected and move. An example is shown in FIG. 6. In this case, additional information about the moved amount is needed. When moving, the last ESD exists, so the information about the last ESD does not need to be encoded. To find the maximum length, think of the first and the last as connecting, and find the maximum length. As shown in FIG. 6, the length of the part becomes 20 + 3 = 23 when the first and the end are connected. In this method, since the first ESD coded is the maximum length, the maximum bit allocation information can be used as in the method of (2). In this method, however, since the maximum bit allocation can be obtained from the first ESD, there is no need to transmit additional information about the maximum bit allocation. Therefore, this method may be moved so that the maximum length comes out first and the encoding is performed in the same manner as in (3). At this time, additional information about MAX_BIT need not be transmitted.

(5) In case of encoding Context_Run using Arithmetic encoding method, bit allocation for encoding Context_Run is determined for each ESD using the above methods. Create a probability table that can represent symbols according to the assigned quota, and use different probability tables according to the quota. In other words, assuming that a bit allocation amount for representing Context_Run is 7 bits, a probability table having 27 symbols is generated, and arithmetic coding is performed using the probability table to encode exceptional sample data.

As described above, the present invention reduces the number of bits allocated for encoding when an ESD occurs during encoding of an enhancement layer in scalable encoding of a binary mask, thereby increasing the efficiency of the higher layer encoding process. There is an effect to increase.

1 is a block diagram of a MPEG-4 VOP video encoder primarily determined by the present International Standardization Organization;

2 is a conceptual diagram of encoding spatial scalability in flexible encoding supported by MPEG-4 of FIG. 1;

FIG. 3 is an example of scan interleaving for flexible encoding of FIG. 1; FIG.

4 is a basic coding unit diagram of exceptional data in flexible encoding;

5 is an example of bit allocation for Context_Run of exceptional sample data according to the present invention.

(a) is an example of bit allocation when encoding Context_Run of E1.

(b) is an example of bit allocation when encoding Context_Run of E2.

(c) is an example of bit allocation in the case of encoding Context_Run of E3.

6 is an example diagram for allocating bits by moving exceptional sample data having a maximum length of Context_Run to the forefront.

<Explanation of symbols for the main parts of the drawings>

20: VOP forming unit 31: Motion estimation unit

32: motion compensation unit 33: subtractor

34: object internal encoding unit 36: old VOP detection unit

37: Shape information encoder 38: Multiplexer

Claims (15)

In the flexible encoding of a binary mask that encodes a higher layer using a base layer, If ESD occurs during the encoding of the enhancement layer, encoding the ESD by reducing the number of bits allocated for encoding the ESD by subtracting the previously generated Context-Run from the range where the current Context_Run may occur. An exceptional sample coding method in a flexible (scalable) encoding of a binary mask. The method of claim 1, An exceptional sample encoding method according to a flexible (scalable) encoding method of a binary mask, wherein the ESD is encoded by using the information on the ESD coded when addressing the location where the ESD occurs. In the flexible encoding of a binary mask that encodes a higher layer using a base layer, In order to reduce the range in which the current ESD may exist by subtracting as much as the length of the existing ESD when the ESD is generated during the encoding of the enhancement layer encoding process, the subtraction is performed. An exceptional sample coding method in a flexible (scalable) encoding of a binary mask characterized by reducing bit allocation. The method of claim 3, wherein When the Context_Run is encoded using arithmetic coding, if a bit allocation amount for representing each Context_Run is determined, a probability table having a symbol that can be represented by the predetermined bit allocation is made, and arithmetic coding is performed using the generated probability table. An exceptional sample coding method for stretching (scalable) coding of a binary mask. In the flexible encoding of a binary mask that encodes a higher layer using a base layer, When ESD occurs during the Enhancement layer encoding process, the order of occurrence of ESD is reversed from left to right in the addressing of the ESD, and the number of bits of the current ESD is subtracted to reduce the range in which the ESD of the current position exists and the minimum bit allocation amount. An exceptional sample encoding method in a flexible (scalable) encoding of a binary mask, wherein the ESD is encoded by including additional information about the ESD. The method of claim 5, wherein When the Context_Run is encoded using arithmetic coding, when a bit allocation amount for representing each Context_Run is determined, a probability table having a symbol that can be represented by a predetermined bit allocation is made, and arithmetic coding is performed using the generated probability table. An exceptional sample coding method for stretching (scalable) coding of a binary mask. In the flexible encoding of a binary mask that encodes a higher layer using a base layer, When ESD occurs during the Enhancement layer encoding process, obtain the longest Context_Run from the location addressing of the ESD, obtain the minimum bit allocation amount to encode the value, and then, whenever the ESD occurs, The exceptional sample encoding method for the flexible (scalable) encoding of a binary mask, wherein the ESD range of the current position is obtained and encoded by comparing the generated bit allocation with the obtained minimum bit allocation. . The method of claim 7, wherein When Context_Run is encoded using arithmetic coding, when a bit allocation for expressing each Context_Run is determined, a probability table having a symbol that can be represented by a predetermined bit allocation is made, and arithmetic coding is performed using the created probability table. Exceptional Sample Encoding Method for Stretchable (Scalable) Encoding of Binary Masks. In the flexible encoding of a binary mask that encodes a higher layer using a base layer, When ESD occurs during the Enhancement layer encoding process, the order of occurrence of ESD is reversed from side to side in the ESD location addressing, the longest context is obtained, and then the position of the previously generated ESD using the reversed ESD data. Exceptional sample coding in binary (scalable) encoding of binary masks, characterized by subtracting the amount of the current position to obtain and encoding the ESD range, and comparing the generated bit allocation with the obtained minimum bit allocation. Way. The method of claim 9, When the Context_Run is encoded using arithmetic coding, when a bit allocation amount for representing each Context_Run is determined, a probability table having a symbol that can be represented by a predetermined bit allocation is made, and arithmetic coding is performed using the generated probability table. An exceptional sample coding method in stretchable (scalable) encoding of a binary mask. In the flexible encoding of a binary mask that encodes a higher layer using a base layer, If ESD occurs during the Enhancement layer encoding process, move the maximum Context_Run first in the ESD location addressing, then subtract the distance from the previously generated ESD when encoding the next ESD. Obtaining, encoding, and comparing the number of bits generated and the number of bits used in the first encoded ESD, short bits are allocated, and the Context_Run is transmitted using the allocated bits (scalable). Exceptional Sample Encoding Method in Encoding. The method of claim 11, When the Context_Run is encoded using arithmetic coding, when a bit allocation amount for representing each Context_Run is determined, a probability table having a symbol that can be represented by a predetermined bit allocation is made, and arithmetic coding is performed using the generated probability table. An exceptional sample coding method in stretchable (scalable) encoding of a binary mask. The method of claim 11, An exceptional sample coding method according to a flexible (scalable) encoding method of a binary mask, wherein the maximum Context_Run is calculated by assuming that the beginning and the end are connected like a circle and adding the lengths of the beginning and the end. In the flexible encoding of a binary mask that encodes a higher layer using a base layer, When ESD occurs during the Enhancement layer encoding process, the order of occurrence of the ESD is reversed to the left and right, and the maximum Context_Run is first come out from the ESD location addressing. An exceptional sample encoding method in a stretchable (scalable) encoding method of a binary mask, characterized in that a range is obtained by subtracting a distance to an ESD. The method of claim 14, When the Context_Run is encoded using arithmetic coding, when a bit allocation amount for representing each Context_Run is determined, a probability table having a symbol that can be represented by a predetermined bit allocation is made, and arithmetic coding is performed using the generated probability table. An exceptional sample coding method in stretchable (scalable) encoding of a binary mask.

Images