KR100559713B1 - Color information encoding / decoding device for parallel scan and its method - Google Patents

Abstract

The present invention relates to a color information encoding / decoding apparatus for a parallel scan for improving coding efficiency and improving a decoded video quality when encoding a video signal for a progressive scan of a video. Performing padding on a macroblock basis when encoding an object unit having a shape, and determining whether to perform frame DCT or field DCT on the macroblock on the object unit and the macroblock on the object boundary after padding; Encoding information indicating the determined DCT type; Shuffling the macro block after determining the frame DCT or the field DCT, and determining the CBPY based on the shuffled shape information after shuffling; After shuffling, color information is encoded by sequentially performing DCT and quantization and performing variable length coding of DCT coefficients. In addition, when the color information is transmitted, information indicating the DCT type is transmitted to the receiver prior to CBPY.

Description

Color information encoding / decoding device for parallel scan and its method

The present invention relates to color / texture information encoding for a progressive scan. In particular, the present invention relates to reducing the number of blocks to be encoded when encoding a video signal for a progressive scan of a video, thereby improving coding efficiency and improving decoded video quality. The present invention relates to a color information encoding / decoding apparatus for a parallel scan and a method thereof.

In general, there are progressive and interlaced methods for transmitting a motion image. In this case, progressive frames are constructed one by one when the frame is constructed. On the other hand, an interlaced frame is constructed by forming two fields one by one in a frame, and then inserting two fields one by one. The height of each field (number of lines) is half the height of the frame. An example illustrating this is FIG. Figure 1 (a) shows a forward scan frame, Figure 1 (b) shows two fields

[Top field and even field] are shown. The odd field and the even field are each arranged one by one in a row (solid arrow for odd field and dotted arrow for even field), and by inserting lines of each configured field (solid arrow and dotted arrow are mixed). A) configures a parallel scan frame. Here, there is a time difference between the two fields when constructing the odd field and the even field. Figure 1 (b) shows the characteristics of such a perturbed scan frame, in which case the odd fields are temporally advanced (in some cases even fields may be temporally advanced).

Due to the time difference between the odd field and the even field, the signal characteristics between adjacent lines in the perforated scan frame may be different. This is especially true for images with a lot of motion. Therefore, when the conventional scanning frame uses an existing video encoding apparatus, for example, motion estimation and compensation, and discrete cosine transform (DCT), the coding efficiency is reduced. Methods to avoid this problem (e.g. field-based motion estimation / compensation, adaptive frame or field DCT) have already been studied, and MPEG- 2, such as a video encoding standard.

Meanwhile, the International Standard on Next-Generation Image and Audio Coding Technology and System Construction, which has been developed and resolved by the MPEG Group, which has developed and decided on international standards (MPEG-1 and MPEG-2) on image and audio coding technology and system configuration, The standard proposal (MPEG-4: ISO / IEC JTC1 / SC29 / WG11 MPEG Phase 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 cannot be supported by existing known standards.

MPEG-4 provides new methods for communication, access, and manipulation of video and audio data (eg, object-oriented interactive functions and access through networks with different characteristics).

In addition, it provides a characteristic that operates usefully in a communication environment where errors are easily generated and a low transmission rate communication environment. In addition, it integrates computer graphics technology to provide 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 multimedia information is possible to support the functions required for all possible low-cost, high-performance applications that are newly developed or will be developed. Among them are the transmission and storage functions and the improved compression efficiency required for cost reduction.

Applications currently expected to apply MPEG-4 technology include interpersonal communication (IPC) such as Internet multimedia (IMM), interactive video games (IVG), video conferencing and video telephony. Networked Database Services (NDB) using Communications, Interactive Storage Media (ISM), Multimedia Mailing (MMM), Wireless Multimedia (WMM), ATM networks, Remote Emergency Systems (RES) and Remote Video Surveillance (RVS).

In order to support existing applications or applications expected in the future, an image encoding technology that can be edited to allow users to communicate, find and read only the objects in the image, and to cut and paste them can be edited.

MPEG-4, a new video and audio encoding technology that is currently under global standardization, is designed to meet this need.

Unlike MPEG-1 and MPEG-2, which are standard video coding schemes, MPEG-4 is in the process of standardizing a method for performing object-based coding or manipulation. To this end, encoding and transmission of shape information representing an object area in a frame together with existing color / texture information is required. Therefore, color coding and shape information coding are required in MPEG-4 which compressively encodes an arbitrarily shaped object. In this case, a phenomenon that existing MPEG-1 and MPEG-2 technologies cannot be considered in the object boundary portion, and new technologies are required to increase coding gain. Currently, new technologies in MPEG-4 include Boundary Block Merging (BBM) and Shape-adaptive Discrete Cosine Transform (SA-DCT).

2 is a configuration diagram of a V-4 video encoder of MPEG-4, which is primarily 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 concepts of Shape Coder and VOP (Video Object Planes) shows the most significant difference. The VOP refers to an object at a point in time on an arbitrary shape of content that can be accessed and edited by the user. The VOP must be encoded for each VOP to support content-based functionality.

When the VOP encoder inputs a VOP for each object image formed in the VOP forming unit 11 to the MOTION ESTIMATION 13, the motion estimating unit 13 performs the macroblock unit from the applied VOP. The motion is estimated.

In addition, the motion information estimated by the motion estimator 13 is input to a motion compensation unit 14 to compensate for the motion. The VOP whose motion is compensated by the motion compensator 14 is input to the subtractor 16 together with the VOP formed by the VOP forming unit 11 to detect a difference value, and the difference value detected by the subtractor 16 is The internal information of the object is encoded by the object internal encoding unit 18 in units of sub blocks of the macro block.

For example, after the X and Y axes of the macroblock 16 * 16 are subdivided into 8x8 subblocks having 8 pixels each of M / 2 x N / 2, the object internal information is encoded.

On the other hand, the VOP whose motion is compensated by the motion compensator 14 and the internal information of the object encoded by the object internal encoder 18 are added to the adder 17, and the output signal of the adder 17 is transferred. The previous VOP which is input to the VEV detector 15 (PREVIOUS RECONSTRUCTED VOP) 15 and is the VOP of the image immediately before the current image is detected.

In addition, the previous VOP detected by the previous VOP detector 15 is input to the motion estimator 13 and the motion compensator 14 and used for motion estimation and motion compensation.

The VOP formed by the VOP forming unit 11 is input to a shape coding unit 12, and shape information is encoded.

Here, the output signal of the shape information encoder 12 may vary depending on the field in which the VOP encoder is applied, and as shown by a dotted line, the output signal of the shape information encoder 12 may be used as the motion estimation unit 13. ), And may be input to the motion compensator 14 and the object internal encoder 18 to be used for encoding motion estimation, motion compensation, and internal information of the object.

In addition, motion information estimated by the motion estimation unit 13, object internal information encoded by the object internal encoding unit 18, and shape information encoded by the shape information encoding unit 12 are applied to the multiplexer 19. After being multiplexed and multiplexed, a bit stream is transmitted to a multiplexer which multiplexes and outputs the outputs of a plurality of encoders, although not shown in the drawing.

In this concept of Moving Picture Expert Group-4 (MPEG-4), one of the biggest features is the shape information encoder. This is the biggest difference that distinguishes MPEG-1 and MPEG-2.

The basic processing unit in video encoding in MPEG-4 is a macro block (MB), and has a size of 16 pixels / line x 16 lines / MB. When encoding an object having arbitrary shape information, MB can be classified into three types in association with shape information.

First, there is a MB (transparent MB) when included outside the object area. In TMB, there is no object region in the MB, and an encoding process is unnecessary.

Secondly, there are MBs (opaque MBs) in the case of being included in the object area, and existing coding methods (eg, MPEG-1 and MPEG-2) can be used directly.

Finally, there is a MB (boundary MB) in which an object region and a non-object region coexist. In order to decode such an MB in a receiver, shape information indicating an object area and color information in the object area must be encoded.

An example is shown in FIG. Here, the hatched parts represent the area of the object to be encoded, and each of the small squares shows MB. In this example, TMB, OMB, and BMB are 6, 12, and 22, respectively.

The configuration diagram of the color information encoding technique of each MB in MPEG-4 is as shown in FIG. Existing technologies largely consist of padding, adaptive frame / field decision, shuffling procedure, discrete cosine transform and quantization (DCT and quantization).

Three codings of the right part of FIG. 4 are steps of encoding information required to be transmitted for color coding. The numbers shown show the order in which they are sent. Briefly, after performing the quantization process, the coefficients are determined to determine the coded block pattern for luminance (CBPY), and the transmission is based on the shape information before shuffling. Only when there is a block to encode through CBPY, dct_type, which is information indicating whether a frame DCT or a field DCT, is transmitted. Thereafter, the quantized DCT coefficients are then variable length coded (VLC). Details and problems of each of the above steps are as follows.

First, in the padding process of FIG. 4, in order to perform 8x8 DCT, a color signal value of an 8x8 block must be defined, and in the case of BMB, there is no color signal value (background rather than an object). To avoid this problem, padding is performed to fill the background part without color information. On the other hand, it is not necessary to perform padding in the TMB that does not include the object region to be encoded or in the OMB having the color information values in all the pixels in the MB. If the BMB is in the inter mode, it fills with a value of "0", and in the intra mode, the low pass extrapolation (LPE) padding is performed. In contrast, the intra mode encodes the signal information in the MB using coding such as DCT, whereas the inter mode predicts the color signal of the current MB from the previous frame and transmits the prediction error.

As shown in Fig. 5, four blocks (block, 8 pixels / row x 8 rows / block) are included in one MB, and each block is named A, B, C, and D for convenience of description. Padding in the conventional technology is performed in units of blocks. In the case of a block in which the object region does not include one pixel, it does not have any color information value even after padding. 6A illustrates an example of BMB, where x is a background pixel (a pixel not included in an object), and o is a pixel included in an object. When the BMB as shown in FIG. 6 (a) is input, if the existing 8x8 unit padding is performed, the result value as shown in FIG. 6 (b) is obtained. In FIG. 6B, p denotes padded pixels, and? Denotes pixels that are not padded and whose color information is not defined.

If the progressive scan frame is encoded, the problem does not occur due to pixels whose color information values are not defined. The transparent block can be known based on the shape information, and the DCT can be performed after padding on a block (non-transparent block) including at least one object region.

However, in the coding scheme of a parallel scan frame, an adaptive frame / field DCT decision process is required. As described below, the determination is made based on the values of all the pixels in the BMB. Unlike the OBM, there is a problem that pixels in which the color value is not defined may occur in the BMB.

Next, look at the process of adaptive frame / field DCT determination and shuffling, similar to the technique in MPEG-2, the existing video coding standard. Compare If the similarity between neighboring lines is large, frame-based DCT is performed. If the similarity is small, field-based DCT is performed. The judgment of the similarity comparison is as follows.

[Equation 1]

i = 06j = 015 (p2i, j-p2i + 1.j) 2+ (p2i + 1, j-p2i + 2, j) 2> i = 06j = 015 (p2i, j-p2i + 2.j) 2 + (p2i + 1, j-p2i + 3, j) 2

In this equation, i and j denote positions of the vertical and horizontal axes of the pixels in MB, respectively. If the left side is larger than the right side as in the above equation, the field DCT is performed, otherwise the frame DCT is performed. Information about this decision (dct_type) is transmitted to the receiver using 1 bit. The MB determined by the field-based DCT performs a shuffling procedure, which is a process of reconfiguring the odd field and the even field to be adjacent to each other by changing a row position as shown in FIG. 7.

In the case of OMB, this decision equation is effective because it is the same as in the case of MPEG-2. However, in the case of BMB generated in MPEG-4, there is a risk that the number of blocks to perform DCT may increase. Consider the case where the MB to be coded is BMB in Fig. 6A. Assuming padding is performed and '?' In FIG. 6B is filled with "0", the result of performing the shuffling procedure is shown in FIG. 8. If it is determined as a frame DCT, padding in units of blocks may be performed, and DCT coding may be performed on the C block and the D block. In the case of FIG. 8, four blocks A, B, C, and D may be used. Should be encoded.

Therefore, in the case of OMB, four blocks are encoded regardless of the field DCT and the frame DCT, so there is no problem. However, in the case of BMB, when only the above equation is applied, there is a problem in that the number of blocks to be encoded increases as in the above example.

In addition, when looking at the discrete cosine transform and quantization (DCT and quantization) process, DCT is performed in 8x8 block units. If it is a transparent block, only the block filled with padding is encoded without performing DCT. After quantizing the DCT coefficients, the variable length coding (VLC) is performed.

The method currently described in MPEG-4 takes shape information before shuffling the determination of whether it is a transparent block. As can be seen from Figure 4, the shape information input in the DCT step is the shape information before shuffling. Therefore, there is a case where color information is not encoded. In the case of FIG. 8, since the determination of the transparent block is judged using FIG. 6A, only two C and D blocks of FIG. 8 are DCT. Therefore, the A and B blocks of FIG. 8 are not transmitted.

Next, referring to the CBPY, dct_type, and DCT coefficient encoding processes, CBPY, dct_type, and DCT coefficients are syntaxes essential for transmitting color information.

After quantizing the DCT coefficients, a block with no quantization coefficients transmitted may occur. Information indicating which block's DCT coefficient is to be transmitted to the receiving end, this information is called CBPY (Coded Block Pattern for Luminance). In conventional MPEG-2, which encodes all blocks in each MB, the number of CBPY cases is 16, so that one variable length code (VLC) table having a length of 16 is used. However, in MPEG-4, which encodes an object having an arbitrary shape, it is possible to know how many blocks are non-transparent blocks out of four luminance blocks in MB, and thus CBPY VLC tables are applied differently. For example, as shown in (a) of FIG. 6, only two blocks may include an object region, and only four cases may occur, and the VLC table length of CBPY is four. The decoder knows the shape information before decoding the color information. Therefore, the VLC table for decoding the CBPY may be selected by determining the decoded shape information.

Therefore, in the conventional method, CBPY is transmitted based on shape information before shuffling as shown in FIG. In case of FIG. 6A, CBPY is transmitted based on two blocks. The decoder decodes CBPY on the basis of two blocks in a state in which dct_type cannot be decoded. Therefore, CBPY and dct_type can be decoded with the same value as the encoder, but the above two blocks of FIG. 6 (b) cannot be transmitted. This causes a disadvantage of seriously degrading the decoded image quality.

That is, FIG. 6A shows an MB in which an object region exists only in two lower blocks C and D among four blocks. If this is padded, the signal is filled only in the lower two blocks as shown in FIG. Assume that the color information of the pixels in the upper two blocks A and B is filled with " 0 " before performing adaptive frame / field determination. If it is determined to perform the field DCT, shuffling is performed, and Fig. 8 shows this shuffling result. In the conventional method, since encoding is based on shape information before the shuffling process, encoding is performed on the lower two blocks C and D. CBPY also uses the VLC table when there are two non-transparent blocks in MB, based on (a) of FIG. Therefore, since the upper two blocks A and B are not transmitted, the quality of the decoded video is severely degraded.

9 (a) shows a BMB having shape information only on the even field side among four blocks A, B, C, and D. FIG. When padding with this BMB, all blocks are filled. With this block, adaptive frame / field DCT decision process is performed. If the field DCT is determined, shuffling is performed, and the result is shown in FIG. Four blocks are quantized with DCT, and CBPY is encoded using the VLC table in the case where four blocks include a shape region. In this case, since only the information (value obtained by performing padding) irrelevant to the original color signal is located in the upper two blocks of FIG. 9 (b), this reduces coding efficiency.

Accordingly, the present invention has been proposed to solve the above-mentioned problems in encoding the color information for the conventional progressive scan, and the present invention is to reduce the number of blocks to be encoded when encoding the video signal for the progressive scan of the video encoding It is an object of the present invention to provide a color information encoding / decoding apparatus and a method for a parallel scan for improving efficiency and improving a decoded image quality.

According to an aspect of the present invention, there is provided a method of encoding color information for performing a parallel scan, comprising: performing padding in units of macroblocks when encoding an image of an object unit having an arbitrary shape; Determining whether to perform a frame DCT or a field DCT on the macro block after performing the padding; Encoding information indicating the determined DCT type; Shuffling the macro block after determining whether to perform the frame DCT or the field DCT; Determining and encoding a CBPY according to the shuffled shape information; And performing DCT and quantization after the shuffling and variable length encoding the DCT coefficients.

After shuffling the shape information of the macroblock determined by the field DCT, the CBPY is encoded by selectively applying a variable length code table based on the number of non-transparent blocks.

In the above, the macroblock on the object boundary is determined to perform a frame DCT.

In the macroblock unit padding, when the macroblock is an inter macroblock, “0” is inserted, and when the macroblock is an intra macroblock, padding is performed by inserting an average value of color information of an object area. It features.

In the above, before determining the frame / field DCT as a discriminant for determining the frame / field DCT, the shape information of the corresponding BMB is shuffled, and the frame DCT or field is determined according to the number of non-transparent blocks of the shuffled macroblock. It is characterized by determining the DCT. The discrimination equation is' i = 06j = 015 (p2i, j-p2i + 1.j) 2+ (p2i + 1, j-p2i + 2, j) 2> i = 06j = 015 (p2i, j-p2i + 2 .j) 2+ (p2i + 1, j-p2i + 3, j) 2 '. I and j denote positions of the vertical and horizontal axes of the pixels in the MB, respectively.

In the above, if the number of non-transparent blocks is increased than the number of non-transparent blocks before shuffling, the frame DCT is determined. If the number of non-transparent blocks is reduced to the number of non-transparent convex before shuffling, Frame / field DCT is determined.

In the above, if the number of non-transparent blocks of macroblocks after shuffling is searched and the number of non-transparent blocks after shuffling is equal to the number of non-transparent blocks before shuffling, the frame / field DCT is determined by the above discriminant. Characterized in determining.

In the above, the shape information of the corresponding BMB is shuffled before the frame / field DCT is determined using the discriminant equation, and then the number of non-transparent blocks of the macroblock and the number of non-transparent blocks of the macroblock before shuffling are determined. In comparison, if the number is different from each other, the frame DCT or the field DCT may be determined using the above-described discriminant.

In the above, the number of non-transparent blocks of macroblocks and the number of non-transparent blocks of macroblocks before shuffling are compared, and if the number is different, unconditionally determining the frame / field DCT toward the smaller number of non-transparent blocks. It features.

In the above, if the non-transparent block of the macroblock is present in only one field after the shuffling, the unit-level DCT is performed unconditionally.

In the above, when the object area and the area where no object exists in the macroblock are BMB, the frame DCT is unconditionally determined.

In the above, when the frame DCT is unconditionally determined for the BMB in which an object region and an objectless region exist simultaneously in the macroblock, information on the DCT type is not transmitted to the decoder.

Another encoding method according to the present invention for achieving the above object,

In the color information encoding method for a progressive scan,

The padding is performed in units of macroblocks, and after determining whether to perform a frame DCT or a field DCT on the macroblock on the object unit and the macroblock on the object boundary after encoding, the information indicating the determined DCT type is preferentially encoded. Transmitting to; Encoding the CBPY obtained by the shuffled shape information after the transmission of the information indicating the DCT type and transmitting the encoded CBPY to the receiving end; And performing variable length encoding on the DCT coefficients after transmitting the CBPY to the receiving end.

In the above, when the encoded macroblock is a macroblock on an object boundary, additional information about an encoding type is not transmitted to the decoder side.

Another encoding method according to the present invention for achieving the above object,

In the color information encoding method for a progressive scan,

The encoding of the field or frame unit for shape information encoding is determined, and the color information is encoded by determining a frame DCT or a field DCT of color information using the determined information.

In the above, when the shape information encoding is a field unit, the color information encoding may be determined in a field unit.

In the above, if the shape information encoding is a frame unit, the color information encoding may be determined in a frame unit.

In the above, when the color information encoding is determined in units of frames, the additional information indicating the DCT type is not transmitted to the decoder.

In the color information encoding, after determining whether the unit of the shape information encoding is a frame unit or a field unit, CBPY is encoded by selectively applying a variable length code table according to the checking result.

Another encoding method according to the present invention for achieving the above object,

In the color information encoding method for a progressive scan,

Performing padding on a macroblock basis when encoding an image of an object unit having an arbitrary shape, and determining whether to perform frame DCT or field DCT on the macroblock on the object unit and the macroblock on the object boundary after the padding; Encoding information indicating the determined DCT type; Shuffling the macroblock after determining the frame DCT or field DCT, and determining the CBPY based on the shuffled shape information after the shuffling; Performing additional padding on an 8 * 8 block containing an object after shuffling; After the additional padding is performed, color information is encoded by performing DCT and quantization and variable length encoding the DCT coefficients.

When the macroblock to be encoded is BMB and the field DCT and the frame DCT are determined and performed, the CBPY is encoded and transmitted by selectively applying a variable length code table based on the shuffled shape information.

In the above, before transmitting the encoded CBPY, additional information indicating a DCT type indicating whether the corresponding macroblock is a frame DCT or a field DCT is first transmitted, and then the CBPY is transmitted.

In the above, when the macroblock of the color information to be encoded is BMB, it is determined to perform the frame DCT unconditionally.

When the macroblock to be encoded is BMB and is unconditionally determined to be a frame DCT, the CBPY is first transmitted to the receiver, and after that, the additional information dct_type indicating the DCT type only when there is a block to be encoded is decoded. Characterized in that the transmission to.

In the above, when the macroblock to be encoded is OMB, the CBPY is encoded and transmitted by selectively applying a variable length code table based on the shuffled shape information.

In the above, before transmitting the encoded CBPY, information indicating a DCT type indicating whether the corresponding macroblock is a frame DCT or a field DCT is first transmitted, and then the CBPY is transmitted.

In the above, when the macroblock of the color information to be encoded is OMB, it is determined to perform the frame DCT unconditionally.

When the macroblock to be encoded is OMB and is unconditionally determined to be a frame DCT, CBPY is first transmitted to the receiving side, and after that, dct_type, which is an additional information indicating the encoding type, is transmitted only when a block to be encoded exists. It is characterized by.

An encoding device according to the present invention for achieving the above object,

In the color information encoding apparatus for a parallel scan,

A padding unit configured to perform padding in units of macro blocks when encoding an image of an object unit having an arbitrary shape; A frame / field determiner configured to determine whether to perform a frame DCT or a field DCT on a macroblock of an object unit output from the padding unit and a macroblock on an object boundary; A DCT type coding unit encoding additional information indicating a DCT type determined by the frame / field determining unit; A shuffling unit for shuffling the macro block output from the frame / field determination unit; A CBPY variable length coding unit for determining CBPY based on the shape information obtained by the shuffling unit and variable length encoding the determined CBPY; A DCT and a quantization unit for DCT and quantizing the signal obtained by the shuffling unit; And a variable length coding unit for variable length coding the DCT coefficients obtained by the DCT and the quantization unit.

An encoding device according to the present invention for achieving the above object,

In the color information encoding apparatus for a parallel scan,

A padding unit configured to perform padding in units of macro blocks when encoding an image of an object unit having an arbitrary shape; A frame / field determiner configured to determine whether to perform a frame DCT or a field DCT on a macroblock of an object unit output from the padding unit and a macroblock on an object boundary; A DCT type coding unit encoding additional information indicating a DCT type determined by the frame / field determining unit; A shuffling unit for shuffling the macro block output from the frame / field determination unit; A CBPY variable length coding unit for determining CBPY based on the shape information obtained by the shuffling unit and variable length encoding the determined CBPY; A field unit padding unit configured to perform field unit padding on an 8 * 8 block including an object output from the shuffling unit; A DCT and quantization unit for DCT and quantizing the signal obtained by the field unit padding unit; And a variable length coding unit for variable length coding the DCT coefficients obtained by the DCT and the quantization unit.

Decoding method according to the present invention for achieving the above object,

Extracting the number of non_transparent blocks from the macroblocks to be decoded of the received video signal, shuffling the macroblocks, and extracting the number of non_transparent blocks from the shuffled macroblocks;

Decoding the received CBPY in which the non_transparent block in the shuffled macroblock exists in only one field and performing field inverse discrete cosine transform;

If the number of non_transparent blocks in the macroconvex before the shuffling and the number of non_transparent blocks in the shuffled macroblock are the same, the received CBPY is decoded, followed by decoding additional information indicating a DCT type, and then selectively according to the decoded DCT type information. Performing field or frame inverse discrete cosine transform;

Decoding the received CBPY and performing a field inverse discrete cosine transform if the number of non_transparent blocks in the macroconvex before the shuffling is greater than the number of non_transparent blocks in the shuffled macroblocks;

And if the number of non_transparent blocks in the macroconvex before the shuffling is smaller than the number of non_transparent blocks in the shuffled macroblocks, decoding the received CBPY and performing frame inverse discrete cosine transform.

Decoding apparatus according to the present invention for achieving the above object,

A first block extracting unit which extracts the number of non_transparent blocks from a macroblock to be encoded of the received video signal;

A shuffling unit which shuffles shape information from a macroblock obtained by the first block extracting unit;

A second block extracting unit extracting the number of non_transparent blocks from the macroblocks shuffled by the shuffling unit;

A comparison unit for comparing the number of non_transparent blocks extracted by the first and second block extractors, respectively, and outputting the result as a signal for determining a field / frame inverse discrete cosine transform;

A CBPY decoder which decodes the CBPY corresponding to the macroblock output from the second block extractor;

A DCT type decoder for decoding a DCT type corresponding to a macroblock output from the second block extractor;

It is composed of an inverse discrete cosine transform and an inverse quantization unit that performs field or frame inverse discrete cosine transform and inverse quantization according to the signals output from the CBPY decoder, the DCT type decoder, and the comparator, respectively. It features.

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

10 is a block diagram of a color information encoding apparatus for a parallel scan according to the present invention.

As shown therein, the padding unit 100 performs padding in units of macroblocks when encoding an image of an object unit having an arbitrary shape, and on the macroblock and object boundary of the object unit output from the padding unit 100. A frame / field determination unit 101 for determining whether to perform a frame DCT or a field DCT with respect to a macroblock, and a DCT type coding unit for encoding additional information indicating a DCT type determined by the frame / field determination unit 101. (102), the shuffling unit 103 for shuffling the macro block output from the frame / field determination unit 101, and the shape information obtained from the shuffling unit 103 to determine the CBPY and determine the determined CBPY. CBPY variable length coding unit 104 for variable length coding, DCT and quantization unit 105 for DCT and quantizing the signal obtained by the shuffling unit 103, and DCT coefficients obtained at the DCT and quantization unit 105 Is composed of a coefficient variable length coding unit 106 for variable length coding.

11 is a block diagram of a color information encoding apparatus for a permutation scan according to an embodiment of the present invention, and includes a block of 8 * 8 units in which an object after shuffling is included between the shuffling unit 103 and the DCT and the quantization unit 105. The field unit padding unit 107 is configured to further include field unit padding.

In Fig. 11, blocks having the same reference numerals as in Fig. 10 perform the same functions and functions.

The color information encoding method by the color information encoding apparatus according to the present invention configured as described above is as follows.

First, in comparison with the conventional color information encoding method for a conventional interlocking scan of FIG. 4, FIG. 10 and FIG. 11 both transmit additional information dct_type indicating a DCT type to a receiver first and then CBPY. It can be seen. In addition, the CBPY transmission may select and encode a VLC table using shape information after performing a shuffling process (not performing in the case of frame DCT).

Hereinafter, the operation of each block will be described in more detail.

First, looking at the padding process in the BMB, conventionally performs padding in units of 8x8 blocks. In the progressive scan image, an adaptive frame / field DCT decision process must be performed and signals of all pixels of the MB must be present. Therefore, the use of the existing technology has a problem that the coding efficiency is lowered because the values of all the pixels in the MB may not be defined.

The proposed invention for solving this problem, in the case of the image for the interlocking scan in the MBB, instead of the existing block-by-block padding is performed in MB unit. In other words, the padding unit 100 pads by inserting "0" in the case of inter MB, and pads by inserting an average value of color information of the object area in case of intra MB.

Next, since the frame / field determination unit 101 considering shape information encodes all blocks in each MB in the conventional adaptive frame / field DCT determination, the number of blocks to be encoded is determined even if it is determined as a field DCT or a frame DCT. same. However, when encoding an object having arbitrary shape information such as MPEG-4, there is a problem in that the number of blocks to be encoded is relatively increased.

The proposed invention for solving this problem uses the characteristics of shape information. Shape information encoding is performed before color information encoding. The decoder also performs shape information decoding before color information decoding. Therefore, the encoder and the decoder have the same shape information before the color information encoding / decoding, and when the shape information is used, additional information transmission is unnecessary.

That is, the shape information of the corresponding BMB is shuffled before the frame / field DCT is determined by using an existing frame / field determination equation. If the number of non-transparent blocks is different before and after shuffling, the DCT is determined as the side (field / frame) where the number of non-transparent blocks is small. In other words, if the number of non-transparent blocks before shuffling is less than the number of non-transparent blocks after shuffling, perform frame DCT; if the number of non-transparent blocks after shuffling is less than the number of non-transparent blocks before shuffling, Perform a DCT. In this case, it is not necessary to transmit dct_type, which is additional information indicating the DCT type, to the decoder side.

Alternatively, the shape information of the corresponding BMB is shuffled before the frame / field DCT is determined using an existing frame / field decision equation. If the number of non-transparent blocks increases after shuffling, the frame DCT is determined without using a decision equation. In this case, it is not necessary to transmit dct_type, which is additional information indicating the DCT type, to the decoder side. In addition, when the number of non-transparent blocks decreases after shuffling, the frame / field DCT is determined using an existing equation for determining the frame / field.

Alternatively, the shape information of the corresponding BMB is shuffled before the frame / field DCT is determined using an existing frame / field decision equation. If the number of non-transparent blocks increases after shuffling, the frame DCT is determined without using a decision equation. If the number of non-transparent blocks decreases after shuffling, the field DCT is determined without using a decision formula for frame / field determination. In this case, it is not necessary to transmit dct_type, which is additional information indicating the DCT type, to the decoder side. In addition, if the number of non-transparent blocks after shuffling is the same as the number of non-transparent blocks before shuffling, the frame / field DCT is determined using an existing equation for determining the frame / field.

Alternatively, the frame DCT is unconditionally performed in the case of the BMB in which the object area and the other area in the MB exist simultaneously. In this case, since the decoder can recognize the reproduced shape information without transmitting the dct_type, the transmission is unnecessary. In the case of OMB in which all pixels in the MB are the object region, the frame DCT and the field DCT are determined by the conventional method, and the dct_type is also transmitted only when the block to be encoded exists.

As another method, in shape information encoding, encoding of a field or a frame unit is determined for a parallel scan frame. In other words, if the shape information encoding type is field unit encoding, the color information encoding degree field DCT is performed. If the shape information encoding type is frame unit encoding, the color information encoding degree frame DCT is performed. Therefore, in this method, it is not necessary to transmit dct_type, which is additional information indicating the DCT type, to the decoder side.

Alternatively, when shuffling, if a non-transparent block exists in only one field, field DCT is unconditionally performed. In this case, it is not necessary to transmit dct_type, which is additional information indicating the DCT type, to the decoder side.

Next, the shuffling process is performed by the same method as the conventional method.

Next, referring to the field-based padding process separately added to FIG. 11, instead of DCT the padded signal of the first stage, after padding the shape information and the color information again, the padding of 8x8 unit is again non- Perform additionally on the transparent block and DCT this signal. The padding method here is consistent with the existing method described above. Therefore, the first stage padding in the case of using this stage is only for adaptive frame / field DCT determination.

If it is assumed that MB has a signal characteristic determined by the field DCT, shuffling and padding before the padding value performed in the first step without shuffling will increase the similarity of the signal and increase the efficiency of the DCT. Can be. In addition, the introduction of this scheme can reduce unnecessary signal transmission, as can be seen in the following description.

Next, the DCT and quantization process is the same as the conventional method.

Next, looking at the CBPY and dct_type transmission method, when performing the selection of frame / field DCT in the BMB as in the conventional method, if the CBPY is transmitted before and the dct_type is transmitted, there is a color signal that does not transmit, so the image quality deteriorates. Results in.

Proposed methods of the present invention to solve this problem are as follows.

First, when determining and performing field DCT and frame DCT in BMB, selection of a variable length code (VLC) table for transmitting CBPY is based on shuffled shape information. This is for transmitting all the color information included in the MB. However, in order to decode the CBPY at the receiving end, it should be able to determine whether shuffled reproduced shape information. Therefore, in the present invention, unlike the conventional scheme, before transmitting CBPY, dct_type indicating whether the corresponding MB is a frame DCT or a field DCT is transmitted first. Even when OMB is input, dct_type is transmitted and CBPY is transmitted.

Secondly, in case of determining and performing field DCT and frame DCT in BMB, selection of VLC table for transmitting CBPY is based on shuffled shape information. This is to transmit all the color information included in the MB. However, in order to decode the CBPY at the receiving end, it should be able to determine whether shuffled reproduced shape information. Therefore, in the present invention, unlike the conventional scheme, before transmitting CBPY, dct_type indicating whether the corresponding MB is a frame DCT or a field DCT is transmitted first. When OMB is input, CBPY and dct_type are sent as in the conventional method.

Third, if the BMB is inputted and unconditionally decides to perform the frame DCT, both the BMB and the OMB send CBPY first, and only if there is a block to encode the dct_type.

Fourth, the first and second methods have a disadvantage in that, even when CBPY = 0 (there is no encoding and no DCT coefficient to be transmitted) in the BMB, dct_type, which is additional information indicating the DCT type, must be transmitted. Therefore, even in the case of BMB, if the number of non-transparent blocks before and after shuffling the shape information is the same, CCTY is first transmitted to the decoder side as in the conventional method, and dct_type, which is an additional information for indicating the DCT type only when CBPY = 0 is not used. Send it. Since the number of non-transparent blocks before and after shuffling shape information is the same, the decoder can decode CBPY without knowing the dct_type.

12 is a block diagram of a color information decoding apparatus for a permutation scan according to the present invention.

As shown therein, a first block extracting unit 108 extracts the number of non_transparent blocks from a macroblock to be encoded of the received video signal; A shuffling unit (109) for shuffling shape information from the macroblock obtained by the first block extracting unit; A second block extracting unit (110) for extracting the number of non_transparent blocks from the macroblocks shuffled by the shuffling unit; A comparator 111 for comparing the number of non_transparent blocks extracted by the first and second block extractors 108 and 110 and outputting the result as a signal for determining a field / frame inverse discrete cosine transform; A CBPY decoder 112 for decoding a CBPY corresponding to a macroblock output from the second block extractor 110; A DCT type decoder 113 for decoding a DCT type corresponding to a macroblock output from the second block extractor 110; Inversely performing inverse discrete cosine transform and field quantization on a macro or frame basis according to signals output from the CBPY decoder 112, the DCT type decoder 113, and the comparator 111, respectively. It is composed of a discrete cosine transform and inverse quantization unit 114.

Referring to the color information decoding apparatus according to the present invention configured as described above will be described in the color information decoding method as follows.

First, the first block extractor 108 extracts the number of non_transparent blocks from a macroblock to be decoded of the received video signal, and the shuffling unit 109 shuffles the shape information of the macroblock. In addition, the first block extractor 110 extracts the number of non_transparent blocks from the macroblocks shuffled by the shuffling unit 109.

The comparison unit 111 compares the number of non_transparent blocks in the shuffled macroblocks with the number of non-transparent blocks before shuffling, and converts the result into the control signal for frame IDCT or field IDCT. It delivers to the inverse quantization unit 114.

In addition, the CBPY decoder 112 decodes the encoded CBPY, and the DCT type decoder 113 decodes the encoded DCT type.

Then, the inverse discrete cosine transform and inverse quantization unit 114 selectively selects the macroblock according to the decoded CBPY, the decoded DCT type information, and the field / frame IDCT information obtained from the comparison unit 111. Discrete cosine transform.

That is, in the method for decoding color information according to the present invention, if a non_transparent block exists in only one field after shuffling, the received CBPY is decoded and subsequently field inverse discrete cosine transform is performed. Further, if the number of non_transparent blocks in the macroconvex before the shuffling and the number of non_transparent blocks in the shuffled macroblock are the same, the received CBPY is decoded, and after decoding additional information indicating a DCT type, the decoded DCT type information is decoded. Optionally, the field or frame inverse discrete cosine transform is performed. In addition, if the number of non_transparent blocks in the macroblock before the shuffling is larger than the number of non_transparent blocks in the shuffled macroblock, the received CBPY is decoded and field inverse discrete cosine transform is performed. In addition, when the number of non_transparent blocks in the macroconvex before the shuffling is smaller than the number of non_transparent blocks in the shuffled macroblock, the received CBPY is decoded and frame inverse discrete cosine transform is performed.

According to another method of decoding color information according to the present invention, if shape information is shuffled, and if the number of non_transparent blocks before and after shuffling is mutually different, dct_type, which is additional information indicating the DCT type, is first decoded, and in the case of field DCT, shape information CBPY is decoded depending on the number of non_transparent blocks after shuffling, and CBPY is decoded depending on the number of non_transparent blocks when not shuffled in the case of frame DCT.

In addition, it is possible to know whether the macroblock is BMB or OMB using the shape information which is already decoded. In the case of BMB, dct_type is first decoded, and shape information is shuffled to determine how many blocks are non_transparent blocks and to decode CBPY.

In case of OMB, CBPY is decoded, and in case CBPY is not 0, dct_type is decoded.

As described above, the present invention performs macroblock-based padding, determines a frame / field DCT for MB on an object boundary, and first transmits dct_type information before transmitting CBPY to the decoder for BMB. By doing so, unnecessary signal transmission can be reduced, and since all color signals can be transmitted, there are a number of advantages that can increase coding efficiency such as to improve image quality.

1 is a flowchart illustrating the transmission of frames and fields on a time axis during general progressive and interlaced operations.

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

3 is a relation diagram between an object area having an arbitrary shape in a macro block MB and BMB, OMB, and TMB;

4 is a flowchart illustrating a method of encoding color information in encoding a conventional perforated scan frame;

5 is a diagram in which one macro block is divided into arbitrary blocks;

FIG. 6 is a diagram for explaining a problem in the case where a conventional block-by-block padding is applied to a BMB; FIG.

FIG. 7 is a diagram illustrating a shuffling process of dividing and arranging MBs formed by a parallel scan frame into odd and even fields; FIG.

8 is a diagram of the results obtained after shuffling FIG. 6B;

9 is a diagram showing a BMB and a result of shuffling the same;

10 is a diagram illustrating an embodiment of a color information encoding apparatus for a progressive scan according to the present invention;

11 is another embodiment of a color information encoding apparatus for a parallel scan according to the present invention;

12 is a block diagram of a color information decoding apparatus for a parallel scan according to the present invention.

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

100: padding section 101: frame / field determination section

102: DCT type coding section 103: shuffling section

104: CBPY variable length coding unit 105: DCT and quantization unit

106: coefficient variable length coding unit 107: field unit padding unit

108,110: first and second block extracting section 109: shuffling section

111: Comparative 112: CBPY decryption unit

113: DCT Decoder

114: Inverse Discrete Cosine Transform and Inverse Quantization

Claims (21)

In the color information encoding method for a progressive scan, Performing padding in units of macro blocks when encoding an image of an object unit having an arbitrary shape; Determining whether to perform a frame DCT or a field DCT on the macro block after performing the padding; Encoding information indicating the determined DCT type; Shuffling the macro block after determining whether to perform the frame DCT or the field DCT; Determining and encoding a CBPY according to the shuffled shape information; And Performing DCT and quantization after the shuffling and variable length encoding the DCT coefficients Color information encoding method for a parallel scan, including; The method of claim 1, After shuffling the shape information of the macroblock determined by the field DCT, the CBPY is encoded by selectively applying a variable length code table based on the number of non-transparent blocks. Color information encoding method for parallel scan. The method of claim 1, Determining to perform a frame DCT for the macroblock on the object boundary Color information encoding method for parallel scan. The method of claim 1, In the macroblock unit padding, when the macroblock is an inter macroblock, “0” is inserted, and when the macroblock is an intra macroblock, padding is performed by inserting an average value of color information of an object area. Characterized Color information encoding method for parallel scan. The method of claim 1, When the macroblock is a BMB in which an object region and an objectless region exist simultaneously in the macroblock, the macroblock is determined as a frame DCT. Color information encoding method for parallel scan. The method of claim 5, When the frame DCT is determined for the BMB, information on the DCT type is not transmitted to the decoder. Color information encoding method for parallel scan. In the color information encoding method for a progressive scan, Performing padding in units of macro blocks when encoding an image of an object unit having an arbitrary shape; Determining whether to perform a frame DCT or a field DCT with respect to the padding macro block; Firstly encoding information indicating the DCT type determined in the step and transmitting the information to a receiving end; Shuffling the macro block after determining whether to perform the frame DCT or the field DCT; Encoding CBPY obtained by the shuffled shape information and transmitting the information indicating the DCT type to the receiving end; Color information encoding method for a parallel scan comprising a. The method of claim 7, wherein If the encoded macroblock is a macroblock on an object boundary, additional information about an encoding type is not transmitted to the decoder. How to send color information for a patient scan. In the color information encoding method for a parallel scan, padding is performed in macroblock units when encoding an image of an object unit having an arbitrary shape, and the frame DCT is applied to the macroblock on the object unit and the macroblock on the object boundary after the padding. Or determining whether to perform the field DCT; Encoding information indicating the determined DCT type; Shuffling the macroblock after determining the frame DCT or field DCT, and determining the CBPY based on the shuffled shape information after the shuffling; Performing additional padding on an 8 * 8 block containing an object after shuffling; And performing color coding by performing DCT and quantization after performing the additional padding and variable length encoding the DCT coefficients. 10. The method of claim 9, wherein when the macroblock to be encoded is BMB and the field DCT and the frame DCT are determined and performed, the CBPY is encoded and transmitted by selectively applying a variable length code table based on the shuffled shape information. A color information encoding method for a parallel scan. 10. The method of claim 9, wherein before transmitting the encoded CBPY, additional information indicating a DCT type indicating whether the macroblock is a frame DCT or a field DCT is first transmitted, and then the CBPY is transmitted. Color information encoding method for the. 10. The method of claim 9, wherein if the macroblock of the color information to be encoded is BMB, the frame information is determined unconditionally to perform a frame DCT. The method according to claim 12, wherein the macroblock to be encoded is BMB, and unconditionally determined as a frame DCT, CBPY is first transmitted to a receiving side, and then additional information indicating a DCT type only when a block to be encoded exists. Color information encoding method for a parallel scan, characterized in that for transmitting the dct_type to the decoder side. 10. The method of claim 9, wherein when the macroblock to be encoded is an OMB, the CBPY is encoded and transmitted by selectively applying a variable length code table based on the shuffled shape information. Encoding Method. 15. The method of claim 14, wherein before transmitting the encoded CBPY, information about a DCT type indicating whether the macroblock is a frame DCT or a field DCT is transmitted first, and then the CBPY is transmitted. Color information encoding method. 10. The method of claim 9, wherein if the macroblock of the color information to be encoded is OMB, the frame DCT is unconditionally determined to be performed. 17. The method of claim 16, wherein if the macroblock to be encoded is OMB and is unconditionally determined to be a frame DCT, CBPY is first transmitted to a receiving side, and then dct_type is transmitted only if a block to be encoded exists. Color information encoding method for a parallel scan. An apparatus for encoding color information for progressive scanning, the apparatus comprising: a padding unit configured to perform padding on a macroblock basis when encoding an image of an object unit having an arbitrary shape; A frame / field determiner configured to determine whether to perform a frame DCT or a field DCT with respect to the macroblock of the object unit output from the padding unit and the macroblock on the object boundary; and to inform the DCT type determined by the frame / field determiner. A DCT type coding unit for encoding additional information; A shuffling unit for shuffling the macro block output from the frame / field determination unit; A CBPY variable length coding unit for determining CBPY based on the shape information obtained by the shuffling unit and variable length encoding the determined CBPY; A DCT and a quantization unit for DCT and quantizing the signal obtained by the shuffling unit; And a variable length coding unit for variable length coding the DCT coefficients obtained by the DCT and the quantization unit. An apparatus for encoding color information for progressive scanning, the apparatus comprising: a padding unit configured to perform padding on a macroblock basis when encoding an image of an object unit having an arbitrary shape; A frame / field determiner configured to determine whether to perform a frame DCT or a field DCT on a macroblock of an object unit output from the padding unit and a macroblock on an object boundary; A DCT type coding unit encoding additional information indicating a DCT type determined by the frame / field determining unit; A shuffling unit for shuffling the macro block output from the frame / field determination unit; A CBPY variable length coding unit for determining CBPY based on the shape information obtained by the shuffling unit and variable length encoding the determined CBPY; A field unit padding unit configured to perform field unit padding on an 8 * 8 block including an object output from the shuffling unit; A DCT and quantization unit for DCT and quantizing the signal obtained by the field unit padding unit; And a variable length coding unit for variable length coding the DCT coefficients obtained by the DCT and the quantization unit. A method of decoding color information for a parallel scan, the method comprising: extracting a number of non_transparent blocks from a macroblock to be decoded of a received video signal, shuffling the macroblocks, and then extracting a number of non_transparent blocks from a shuffled macroblock; ; Decoding the received CBPY in which the non_transparent block in the shuffled macroblock exists in only one field and performing field inverse discrete cosine transform; If the number of non_transparent blocks in the macroconvex before the shuffling and the number of non_transparent blocks in the shuffled macroblock are the same, the received CBPY is decoded, followed by decoding additional information indicating a DCT type, and then selectively according to the decoded DCT type information. Performing field or frame inverse discrete cosine transform; Decoding the received CBPY and performing a field inverse discrete cosine transform if the number of non_transparent blocks in the macroconvex before the shuffling is greater than the number of non_transparent blocks in the shuffled macroblocks; Decoding the received CBPY and performing a frame inverse discrete cosine transform if the number of non_transparent blocks in the macroconvex before the shuffling is smaller than the number of non_transparent blocks in the shuffled macroblock. Information Decryption Method. An apparatus for decoding color information for progressive scanning, comprising: a first block extracting unit for extracting a number of non_transparent blocks from a macroblock to be encoded of a received video signal; A shuffling unit which shuffles shape information from a macroblock obtained by the first block extracting unit; A second block extracting unit extracting the number of non_transparent blocks from the macroblocks shuffled by the shuffling unit; A comparison unit for comparing the number of non_transparent blocks extracted by the first and second block extractors, respectively, and outputting the result as a signal for determining a field / frame inverse discrete cosine transform; A CBPY decoder which decodes the CBPY corresponding to the macroblock output from the second block extractor; A DCT type decoder for decoding a DCT type corresponding to a macroblock output from the second block extractor; An inverse discrete cosine transform and inverse quantization unit for performing inverse discrete cosine transform and inverse quantization on a macroblock or field basis according to signals output from the CBPY decoder, the DCT type decoder, and the comparator, respectively; Color information decoding apparatus for a percussion scan, characterized in that configured.