KR100525563B1 - Image information encoding / decoding apparatus and method thereof - Google Patents

Abstract

By applying the block merging technique, the DC coefficient and the AC coefficient of a block which is an out-of-object block in the object boundary block are set by the merging information, and the DC coefficient and the AC coefficient of the neighboring block are predicted using the set coefficients. / Decoding apparatus and method therefor. The image information encoding apparatus sets values for the DC component and the AC components according to the merging information and uses the DC coefficient and the AC coefficient of the neighboring block for predictive encoding. The image information decoding apparatus sets values for the DC component and the AC components according to the merging information and uses the DC coefficient and the AC coefficient of the neighboring block for predictive decoding. As a method, after performing an object merging technique, a block that has been changed into an out-of-object block by merging is DCT-transformed into a block including a signal combined by merging without performing prediction, and DC prediction is performed using the DC value. Also, the block changed into the out-of-object block is not predicted and the quantized DCT coefficients (DC and AC coefficients) of the merged block are copied to the current sub-block for AC / DC prediction. Another method is to copy only the DC coefficient among the quantized DCT coefficients of the merged block and set the AC coefficient to zero. Another method is to set the DC coefficient to 2 ^{(bits_per_pixel-1)} and set the AC coefficient to zero. Another method is to copy the AC coefficient and the DC coefficient of the block to the current sub-block if there is a block other than the out-of-object block in the left block and the upper block.

Description

Image information encoding / decoding apparatus and method thereof

More particularly, the present invention relates to a method and apparatus for encoding and decoding image information, and more particularly, to a method and apparatus for encoding and decoding image information, And a method for encoding and decoding an image signal.

If each pixel of a video signal is represented by digital data of a predetermined bit (for example, 8 bits, 16 bits, etc.), the amount of the information becomes excessively large so that it can be transmitted and received using a communication network or stored in a storage medium such as a CD-ROM The cost is high and the image quality is deteriorated and it is impossible to communicate with the conventional public communication network or the like. Therefore, there is a need to compress image information in order to communicate image information using a conventional communication network and to reduce costs.

In order to compress the image information, the image information is divided into layers as shown in FIGS. 1 (A) - (F). 1 (A) shows a video sequence layer and is made up of a plurality of picture groups (GOPs). One GOP is composed of an I picture, a B picture, and a P picture (see Fig. 1 (B)). One picture is composed of a plurality of slices (see FIG. 1 (C)). One slice is composed of a plurality of macroblocks as shown in FIG. 1 (D) Is composed of 16 x 16 pixels (Pixels). One macro block is composed of four luminance subblocks B1-B4 and eight color difference signal subblocks B5-B12 as shown in FIG. 1 (E), and one subblock is composed of 8 × 8 pixels (See Fig. 1 (F)).

FIG. 2 shows an example of a picture showing an arbitrary object having a macroblock (16 × 16 pixels) for image encoding. As can be seen in Fig. 2, there are three types of macroblocks. That is, three types of object macroblocks 22 consisting of only information in the object, out-of-object macroblocks 23 having no information in the object, and object boundary macroblocks 21 partially having information in the object to be.

The image information encoding technique is applied to the macro block thus constructed. When the original video signal is not an error signal, the 8 × 8 block is DCT-transformed, DC coefficient is quantized, and DC prediction is performed. As shown in FIG. 3, there is a DC prediction method for performing lossless coding by applying a peripheral block and a DPCM in a zigzag order. Here, the white block is an intra-object block, the gray block is an object boundary block, and the black block is an out-of-object block.

The quantized DC coefficient DC_B4 of the subblock B4 of the neighboring macroblock MB1 with respect to the block B of the current macroblock MB2 to be coded for encoding the DC coefficient of the current macroblock MB2 The predicted value DC_P of the block B is set.

The previous block A, the upper left block B and the upper block C of the current block X to be coded are selected as shown in FIG. 4 and the DC coefficient of the previous block A and the upper left block B When the absolute value of the vertical gradient is smaller than the absolute value of the horizontal slope by comparing the absolute value of the vertical gradients and the absolute value of the horizontal slope of the upper left block B and the DC coefficient of the upper block C, A is a DC predictive value for DC coefficient encoding of the current block X. When the absolute value of the horizontal slope is smaller than the absolute value of the vertical slope, the DC coefficient of the upper block C is set to the current block X) of the current block X by subtracting the DC predicted value from the DC coefficient value of the current block X, as a DC coefficient value for encoding the DC coefficient value of the current block X.

As shown in FIG. 5, the prediction of the AC coefficient is performed based on the first row coefficient (when predicted from the upper block C) or the first column coefficient (predicted from the left block A) of the previous block in the same direction as the DC prediction direction determined in FIG. Is used as the predicted values of the first row coefficient or first column coefficient of the current block (X) to perform differential coding.

When predicting the DC coefficient and the AC coefficient as described above, only the sub-block and the object boundary sub-block in the macroblock are scanned and predicted and the sub-block outside the object is not coded.

In each object boundary macroblock, there are three kinds of sub-blocks, i.e., an object sub-block 61, an out-of-object sub-block 63, and an object boundary sub-block 62 and 64 as shown in FIG. The two object boundary subblocks 62 and 64 can be merged with each other, and the encoding efficiency can be enhanced by merging and performing the encoding process.

However, when the Boundary Block Merge (BBM) technique is applied, a merged and merged sub-block, a block made up of sub-blocks outside the object due to the merge process, and a sub- . When the object boundary block merging technique is applied, for example, block 4 of FIG. 7A is changed from an object boundary block to an outside-object subblock through an object boundary block merging process as shown in FIG. 7B.

When the object boundary block merging technique is applied, there exists a block that changes from an object boundary block to an out-of-object block by an object merging process. When applying the BBM technique and predicting the DC coefficient and the AC coefficient, it is not defined how to process a block with an out-of-body block by the merging process. If the block is treated as an object boundary block, unnecessary DC conversion process is performed The transmission of the converted data not only causes an error but also degrades the image quality.

The present invention has been made to solve the above problems,

The object of the present invention is to apply an object boundary block merging technique to macroblocks of image information, and to apply DC coefficient and AC coefficient of a block which is an out-of-object block in an object boundary block to a block An image information encoding / decoding device for encoding and decoding image information by copying and setting the DC coefficient and the AC coefficient of the block and predicting the DC coefficient and the AC coefficient of the neighboring sub-block using the set coefficient, .

Another object of the present invention is to set the DC coefficient and the AC coefficient of a block which becomes an out-of-object block in an object boundary block by the merging information by applying the object boundary block merging technique, And an image information encoding / decoding device for encoding and decoding image information by predicting an AC coefficient and a method thereof.

According to an aspect of the present invention, there is provided an apparatus for encoding an object, the apparatus comprising: an image signal encoding unit for receiving an image signal and comparing object values of corresponding blocks to determine an object boundary block, A block merging unit; A discrete cosine transform unit for receiving a video signal and performing discrete cosine transform; A quantization unit for receiving the quantized transform coefficients output from the DCT unit and quantizing the transform coefficients; A transform coefficient and an encoding block pattern encoding unit for encoding quantized transform coefficients output from the quantization unit and merging information output from the object boundary block merging unit to output texture information; An inverse quantization unit for receiving the transform coefficients quantized by the quantization unit, dequantizing the transform coefficients, and extracting and outputting the transform coefficients; An inverse discrete cosine transform unit for performing inverse discrete cosine transform on the transform coefficients output from the inverse quantization unit; And an object boundary block separator for outputting decoded data by receiving the merging information in the object boundary block merging unit and the video signal output from the inverse discrete cosine transform unit.

An object boundary block padding unit for receiving an image signal and original shape information from the image signal encoding unit of the encoding apparatus and filtering out a non-object part in the sub-block with an average value of the object part or setting the value to zero; An object boundary block merging unit that receives the image signal output from the object boundary block padding unit and the reproduced shape information and performs an object boundary block merging process; A discrete cosine transformation unit that receives the image signal output from the object boundary block merging unit and performs discrete cosine transformation; A quantization unit for receiving the quantized transform coefficients output from the DCT unit and quantizing the transform coefficients; A transform coefficient and an encoding block pattern encoding unit for encoding quantized transform coefficients output from the quantization unit and merging information output from the object boundary block merging unit to output texture information; An inverse quantization unit for receiving the transform coefficients quantized by the quantization unit, dequantizing the transform coefficients, and extracting and outputting the transform coefficients; An inverse discrete cosine transform unit for performing inverse discrete cosine transform on the transform coefficients output from the inverse quantization unit; And an object boundary block separator for receiving the image signal output from the IDCT and the decoded shape information and outputting the decoded data.

The image signal reconstruction unit of the decoding apparatus receives the motion information input from the motion compensation unit and the decoded image information input from the image signal decoding unit to perform a prediction process to decode and output the image information, Wealth; An inverse quantization unit for inverse-quantizing the transform coefficients and the image information decoded and output by the encoding pattern decoding unit by a quantization coefficient; An inverse quantization unit that inversely transforms the inversely quantized transform coefficients output from the inverse quantization unit and outputs the transformed inverse transformed image data as image data in a spatial domain; And an object boundary block separator for outputting the decoded video data in units of macroblocks received from the video signal output from the IDCT unit to a video display unit.

In the encoding and decoding method according to the present invention, after performing the object merging technique, a block transformed into an out-of-object block by merging is DCT-transformed into a block including a signal combined by merging without prediction, Make predictions.

In the other method, when the object boundary block merging technique is applied and AC / DC predictive coding is applied, a block changed from the object boundary block to the object boundary block by the object boundary block merging technique is treated as an out- , And DC and AC components with arbitrary values for AC / DC predictive coding.

Another method is to compare the DC coefficient and the AC coefficients of the quantized DCT coefficients of the merged block to the current sub-block, if the merging information indicates that the current sub-block is a block changed from the object boundary block to the out- Copy to block.

Another method is to copy only the DC coefficient of the quantized DCT coefficients of the merged block to the current subblock if the merging information indicates that the current subblock is a block that has been changed from the object boundary block to the object outside block by the merging process And the AC coefficient is set to zero.

In another method, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-of-object block by the merging process, the DC coefficient is not DCT encoded like the original out- 2 (bits_per_pixel-1) and the AC coefficient is set to zero.

In another method, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-of-object block by the merging process, the out-of-object block in the left block and the upper block, , The AC coefficient and the DC coefficient of the block are copied to the current sub-block.

Hereinafter, the technical idea of the present invention will be described with reference to the drawings.

FIG. 8 is a block diagram showing a configuration of an embodiment of an image information encoder according to the present invention. An object image signal is input from a camera (not shown) to a motion estimation (motion estimator) unit 81 on a macroblock-by-macroblock basis, and motion is estimated in units of macroblocks. The motion information of the macroblock estimated by the motion estimation unit 81 is inputted to a motion compensation unit 82 to compensate the motion. The motion compensated video signal of the macroblock unit in the motion compensation unit 82 is input to the subtractor 83 together with the object video signal of the macroblock unit to detect a difference value, The difference value is input to the image signal encoding unit 84 and the image signal of the object is encoded in units of subblocks. For example, the video signal encoding unit 84 subdivides the X-axis and Y-axis of the macroblock into 8 × 8 sub-blocks each having eight pixels, and then encodes the video signal of the object.

The motion compensated video signal of the macroblock unit and the encoded information of the object encoded by the video signal encoding unit 84 are inputted to the adder 85 and are added together, The output signal is input to the previous video signal detection unit 86 to detect a previous video signal. The video signal of the previous macroblock detected by the previous video signal detecting unit 86 is input to the motion estimating unit 81 and the motion compensating unit 82 to be used for motion estimation and motion compensation.

The motion information estimated by the motion estimation unit 81 and the internal information of the object encoded by the video signal encoding unit 84 are multiplexed by the multiplexing unit 88 and transmitted as a bit stream through the buffer unit 89.

In an embodiment of the present invention, an image signal encoding unit 84 is configured to perform a merging process based on a macroblock and then encode the image signal, and an embodiment of a block diagram showing the configuration is shown in FIG. 9 (A).

The video signal encoding unit of the present embodiment includes a macroblock scanning unit 91 that scans a macroblock in units of subblocks and scans image information, and a subblock that is scanned by the macroblock scanning unit 91 when a merging condition is satisfied A merging process performing unit 92 in the merging macroblock and an image information encoding unit 93 for encoding the merged block in the merging process performing unit 92 in the macroblock.

Here, the merge condition will be described with reference to Figs. 7 (A) and 7 (B). The macroblock scanning unit 91 divides the macroblock into block 1-block 4 as shown in FIGS. 7A and 7B, and stores the outline information of the object while scanning each sub-block. For example, the outline information compares each pixel value with a threshold value, and if it is smaller, it is set to 0, and if it is set to 1, contour information is generated. According to the stored outline information, the merging process performing unit 92 searches the boundary blocks and compares the outline information of the corresponding coordinates between the boundary blocks to perform the merging process. That is, if block 1 and block 4 are object boundary blocks, and block 4 is rotated by 180 ° and there is no pixel superimposed between block 1, object and shape information rotated by block 4 are merged into block 1, Block 4 is changed to an out-of-object block as shown in Fig. The image encoding unit 93 encodes the merged block.

As another embodiment, the video signal encoding unit 84 is configured as shown in Fig. 9 (B). In the independent encoding mode of motion information and image information, the merging process is performed over the entire area of the image without being restricted to the macroblock. Here, the image signal encoding unit 84 includes a full image merging unit 94 for performing a merging process in the entire image, an image encoding unit 95 for encoding the merged and outputted image signal in the full image merging unit 94, ).

The DC coefficient and the AC coefficient prediction are performed by the image signal coding unit 84. When the object boundary merging technique is applied, the merged signal and the block type are received without receiving the existing image signal.

FIG. 10 illustrates a scanning method at the time of predicting the DC coefficient and the AC coefficient after applying the object boundary block merging technique according to the present invention. The white block is an intra-object block, the gray block is an object boundary block, Is an out-of-body block. As shown in FIG. 10, the fourth block of the macroblock MB2, which is the object boundary block in FIG. 3, is changed into an out-of-object block. As can be seen from the dotted line indicating the prediction order, In the method, it is not predicted.

FIG. 11 is a block diagram showing a configuration of still another embodiment of a video signal encoding unit according to the present invention for performing DC coefficient and AC coefficient prediction after merging.

The image signal encoding unit 84 according to the present embodiment receives an image signal and compares the values of pixels corresponding to the respective sub-blocks to determine whether or not the image is an object boundary block, and performs an object boundary block merging process An object boundary block merging unit 111 for performing a merging process on the input image signal, a DCT unit 112 for performing a DCT on the image signal subjected to the merging process in the object boundary block merging unit 111, A quantization unit 113 for receiving and quantizing the transform coefficients and outputting the quantized transform coefficients and the merging information output from the object boundary block merging unit 111, An inverse quantization unit 115 for receiving the inverse quantized transform coefficients received from the quantization unit 113, extracting the transform coefficients, and outputting the transform coefficients; An inverse DCT unit 116 for performing inverse DCT on the transform coefficients output from the base quantization unit 115, and an inverse discrete cosine transform unit 116 for transforming the image signal output from the inverse DCT unit 116, And an object boundary block separator for receiving the merging information from the merging unit 111 and outputting the decoded data.

12 is a block diagram showing a configuration of a video information decoder according to the present invention.

The multiplexed video signal information and motion information input from the coding unit are input to the demultiplexer 111, and the multiplexed signals are separated and output. The motion information output from the demultiplexer 111 is input to the motion decoding unit 122 to detect motion information. The video signal information is input to the video signal decoding unit 125. The video signal decoding unit 125 compares the values of corresponding pixels between the respective blocks to determine whether they are object boundary blocks and performs an object boundary block division process Performs a prediction process, and decodes it to output a video signal. Here, the video signal decoding unit 125 divides the macroblock into block 1-block 4 as shown in FIGS. 7A and 7B, compares the values of corresponding pixels between the sub-blocks, And performs a dividing process. That is, if block 1 and block 4 are object boundary blocks, and there is no pixel superimposed between block 1 and block 1, object and shape information rotated in block 1 in block 1 is divided, and the divided information is again divided 180 The block 4 is constituted. The division process between block 1 and block 2, block 1 and block 3, block 2 and block 3, block 2 and block 4, and block 3 and block 4 can be similarly performed.

The motion compensation unit 123 receives the motion information detected by the motion decoding unit 122 and the image data from the image storage unit 126 to compensate for the motion. The image reconstruction unit 124 receives the motion compensation information output from the motion compensation unit 123 and the image signal output from the image signal decoding unit 125 to reproduce the image and output the image to a video display unit Output.

13 is a block diagram showing a configuration of a video signal decoding unit of a decoder for decoding image information encoded by a merging and predicting process.

The transform coefficient and encoding pattern decoding unit 131 receives the decoded image information input from the demultiplexer 121, compares the values of corresponding pixels between the blocks, determines whether the block is an object boundary block, and performs an object boundary block division process And decodes and decodes the image information by performing a prediction process. The inverse quantization unit 132 multiplies the image information decoded and output by the conversion coefficient and coding pattern decoding unit 131 by a quantization coefficient, and dequantizes the image information. The inverse DCT unit 133 inversely transforms the inversely quantized transform coefficients output from the inverse quantization unit 132 and outputs the inverse transformed DCT coefficients as image data of a spatial range. The object boundary block separator 134 receives the video signal output from the IDCT 133 and outputs the decoded video data in units of macroblocks to a video display unit (e.g., a monitor, a TV, or the like).

As another embodiment, the concept of a VOP (VOP) may be introduced as a unit block having arbitrary shape information in order to compress image information.

The VOP is defined as a rectangle that separates an applied image into a background image and each object image, and includes the separated background image and object image. A predetermined object or a region of an object composed of a predetermined region If so, the image of the object is divided into respective VOPs, and the separated VOPs can be separately encoded.

FIG. 2 is an example of a VOP having an arbitrary shape reconstructed as a macro block (16 × 16 pixels) for image encoding. In the reconstructed VOP, as shown in the drawing, three kinds of macro blocks There are macroblocks in the object composed only of the information in the object, macroblocks in the object not having any information in the object, and object boundary macroblocks partially having the information in the object.

As shown in FIG. 6, there are three kinds of sub-blocks in the object boundary macroblock 21 constituting the VOP, and encoding efficiency can be improved by merging the object boundary sub-blocks and performing the encoding process.

14 is a block diagram showing a configuration of a VOP encoder according to the present invention. A VOP for each object image formed in the VOP forming unit 140 is input to a motion estimation unit 141 to estimate motion in units of a macroblock.

The motion information of the macroblock estimated by the motion estimation unit 141 is inputted to a motion compensation unit 142 to compensate for motion. The motion compensated VOP of the macroblock unit in the motion compensation unit 142 is input to the subtracter 143 together with the VOP formed in the VOP forming unit 140 to detect a difference value, The difference value is input to the image signal encoding unit 144 together with the shape information input from the shape information encoding unit 147, and the image signal of the object is encoded in units of subblocks by the merging process and the DC coefficient and AC coefficient prediction process .

The motion compensated VOP of the macroblock unit and the encoded information of the object encoded by the image signal encoding unit 144 are inputted to the adder 145 and are added together and the output of the adder 145 The signal is input to the previous VOP detection unit 146 to detect the VOP of the previous screen. The VOP of the previous screen detected by the previous VOP detection unit 146 is input to the motion estimation unit 141 and the motion compensation unit 142 to be used for motion estimation and motion compensation. The VOP formed in the VOP forming unit 140 is input to the shape information encoding unit 147 to encode shape information.

The output signal of the shape information encoding unit 147 is determined depending on the field to which the VOP encoding unit is applied. The output signal of the shape information encoding unit 147 is output to the motion estimation unit 141, The motion compensation unit 142, and the image signal encoding unit 144 so as to encode motion estimation, motion compensation, and internal information of the object.

The motion information estimated by the motion estimation unit 141, the internal information of the object encoded by the image signal encoding unit 144, and the shape information encoded by the shape information encoding unit 147 are multiplexed by the multiplexer 148, And is transmitted as a bit stream through the buffer unit 149.

The DC coefficient and the AC coefficient prediction are performed in the image signal coding unit 144. When the object boundary merging technique is applied, the object boundary block merging unit 152 receives the existing image signal, Lt; / RTI >

FIG. 15 is a block diagram showing a configuration of another embodiment of a video signal encoding unit according to the present invention for performing DC coefficient and AC coefficient prediction after merging.

The image signal encoding unit 144 of the VOP encoder according to the present invention receives the image signal and the original shape information and filters the non-object part in the sub-block (i.e., 8x8 block) with the average value of the object part, An object boundary block merging unit 152 that receives the image signal and the reproduced shape information output from the object boundary block padding unit 151 and performs an object boundary block merging process, A discrete cosine transformation unit 153 that receives the image signal output from the object boundary block merging unit 152 to perform discrete cosine transform (DCT), and a transform coefficient output unit 153 that receives the transform coefficient output from the DCT unit 153 A quantization unit 154 for quantizing and outputting the quantized transform coefficient and the merged information output from the object boundary block merging unit 152, An inverse quantization unit 156 for receiving the transform coefficients quantized by the quantization unit 154 and dequantizing the transform coefficients to extract transform coefficients and outputting the transform coefficients; An inverse discrete cosine transform unit 157 that performs inverse discrete cosine transform on the transform coefficients output from the inverse discrete cosine transform unit 157 and outputs the decoded transformed image signal and the reconstructed shape information, And an object boundary block separator 158 for outputting VOP data in units of macroblocks.

16 is a block diagram showing the configuration of still another embodiment of the VOP decoder according to the present invention.

The multiplexed image information output from the VOP coding unit and transmitted to the receiving side is input to the demultiplexer 161, and the multiplexed signals are separated and output. The image information output from the demultiplexer 161 is decoded by the shape information decoding unit 162 and output. The motion information output from the demultiplexer 161 and the shape information output from the shape information decoding unit 162 are input to the motion information decoding unit 163 to detect motion information. The video signal information output from the demultiplexer 161 and the shape information output from the shape information decoding unit 161 are input to the video signal decoding unit 166. The video signal decoding unit 166 decodes the merged blocks And decodes and outputs the video signal. The motion compensation unit 164 receives the motion information detected by the motion information decoding unit 163, the shape information output from the shape information decoding unit 162, and the VOP data input from the VOP storage unit 167, do. The VOP reconstruction unit 165 receives the shape information output from the shape information decoding unit 162, the motion compensation information output from the motion compensation unit 164, and the video signal output from the video signal decoding unit 166, Reproduce. The combining unit 168 receives the VOP signal output from the VOP reproducing unit 155 and combines the VOP signals.

FIG. 17 is a block diagram showing a configuration of a VOP reconstruction unit of a VOP decoder that decodes image information encoded by a merging and predicting process.

The merging information extracting unit 175 receives the shape information output from the shape information decoding unit 162 and extracts the merging information by the same merging process as the encoding unit and outputs the merging information to the transform coefficient and encoding pattern decoding unit 171, And outputs it to the separator 174. The conversion coefficient and coding pattern decoding unit 171 receives the merging information output from the merging information extracting unit 175 and the video signal output from the demultiplexer 161 and divides the merged block and performs a prediction process Decodes and outputs the image information. The inverse quantization unit 172 dequantizes the transform coefficients and the image information decoded by the encoding pattern decoding unit 171 by a quantization coefficient. The inverse discrete cosine transformation unit 173 inversely transforms the inversely quantized transform coefficients output from the inverse quantization unit 172 and outputs the inverse transformed image data as spatial data. The object boundary block separator 174 receives the video signal output from the IDCT 173 and the merging information output from the merging information extractor 175 and outputs the decoded VOP data for each macroblock do.

The object boundary block merging process and the DC and AC coefficient prediction process are performed in the video signal encoding units 84 and 144 in the encoder and in the video reproducing unit 124 and the VOP reproducing unit 165 in the decoder.

When the transformation coefficients and the encoding pattern coding units 114 and 155 generate texture information, the merging information output from the object boundary block merging units 111 and 152 is merged with the current subblock (DC and AC coefficients) of the merged block so that the current sub-block is not DCTed in the DCT units 112 and 153, and the quantized DCT coefficients (DC and AC coefficients) Is copied to the current sub-block and used for the prediction process.

For example, in FIG. 3 and FIG. 10, since the block 4 of the macroblock MB2 has been changed from the object boundary block to the object-outside block by the merging process, in block 2 of the macroblock MB2 including the merged information DCT coefficients are copied and the block 1 of the macroblock MB3 is predicted using the values.

4, when the DC coefficient for the current block X is to be subjected to under-coding, the absolute value of the vertical gradient of the DC coefficient of the previous block A and the upper-left block B, When the absolute value of the vertical slope is smaller than the absolute value of the horizontal slope by comparing the absolute value of the horizontal slope of the DC coefficient of the block B with that of the upper block C, And when the absolute value of the horizontal slope is smaller than the absolute value of the vertical slope, the DC coefficient of the left block (A) is used as the DC value for coding the DC coefficient of the current block (X) The DC coefficient of the current block X is predictively coded as a predictive value.

At this time, if the blocks (blocks A, B, and C) used for DC coefficient prediction coding are changed from the object boundary block to the out-of-body block by the object boundary block merging process, And uses the information after copying the DCT coefficients of the coded blocks.

5, the first row coefficient of the left block A or the first row coefficient of the upper block C used for predicting the DC coefficient determined in the method described with reference to FIG. It is used to predict the first dimension or the column dimension. At this time, if the blocks (blocks A and C) used for AC coefficient prediction coding are changed from the object boundary block to the out-of-body block by the object boundary block merging process, the information of the block in the macroblock including the block And uses the information after copying the DCT coefficients of the coded blocks.

FIG. 18 is a flowchart showing a coding method for performing another merging prediction process according to the present invention.

The merging information output from the object boundary block merging unit 111 (the object boundary block merging unit 152 in the case of the video signal encoding unit 144 of the VOP encoder) constituting the video signal encoding unit 84 of the encoder in step S181, . In step S182, it is determined whether block 1 has changed from the object boundary block to the object outside block from the merging information. If block 1 has changed from an object boundary block to an out-of-object block, the quantized DCT coefficient value of the block obtained by merging the information of block 1 in step S183 using the merging information is copied to block 1, and the process proceeds to step S184 do. If the block 1 does not change from the object boundary block to the object outside block, the process also proceeds to step S184.

In step S184, it is determined whether block 2 has changed from the object boundary block to the object outside block from the merging information. If block 2 has changed from an object boundary block to an out-of-object block, the merged information is used to copy the quantized DCT coefficient value of the block obtained by merging the information of block 2 in block S185 into block 2, and proceeds to step S186 do. If the block 2 does not change from the object boundary block to the object outside block, the process also proceeds to step S186.

In step S186, it is determined whether block 3 has changed from the object boundary block to the object outside block from the merging information. If the block 3 has changed from an object boundary block to an out-of-object block, the merged information is used to find the block in which the information of the block 3 is merged in step S187, the quantized DCT coefficient value of the block is copied to the block 3, The process proceeds to step S188. If the block 3 does not change from the object boundary block to the out-of-object block, the process also proceeds to step S188.

In step S188, it is determined whether block 4 has changed from the object boundary block to the object outside block from the merging information. If block 4 has changed from an object boundary block to an out-of-object block, in step S189, the merged information is used to find the block encoded by merging the information of block 4, and the quantized DCT coefficient value of the block is copied to block 4 , The flow advances to step S190. If the block 4 does not change from the object boundary block to the object outside block, the process also proceeds to step S190.

In step S190, prediction of the DC coefficient and the AC coefficient is performed. It is determined whether the current macroblock encoded in step S191 is the last macroblock including the video information. If the current macroblock is not the last macroblock including the video information, the process returns to step S181 to perform the following process. If it is a macroblock, terminate the program.

FIG. 19 is a flowchart illustrating a coding method for performing a post-merging prediction process according to the present invention. Block 1 is an object boundary block, because it does not change into an out-of-body block by merging the object boundary.

The merging information output from the object boundary block merging unit 111 (the object boundary block merging unit 152 in the case of the video signal encoding unit 144 of the VOP encoder) constituting the video signal encoding unit 84 of the encoder in step S201, . In step S202, it is determined whether block 2 has changed from the object boundary block to the object outside block from the merging information. If block 2 has changed from an object boundary block to an out-of-object block, the quantized DCT coefficient value of block 1 is copied to the DCT coefficient of block 2 in step S203, and the process proceeds to step S204.

If block 2 does not change from an object boundary block to an out-of-object block, it is determined in step S204 whether block 3 has changed from the merging information to an out-of-object block. If block 3 has changed from an object boundary block to an out-of-object block, it is determined in step S205 whether the merger is a vertical merger. In the case of vertical merging, the quantized coefficient value of block 1 is copied to block 3 in step S207, and the process proceeds to step S208. If it is not a vertical merging, the quantized coefficient value of block 2 is copied to block 3 in step S206, and the process proceeds to step S208.

If block 3 has not changed from an object boundary block to an out-of-object block, it is determined in step S208 whether block 4 has changed from an object boundary block to an out-of-matter block. If block 4 has changed from an object boundary block to an out-of-object block, it is determined in step S209 whether the merger is a horizontal merger. In the case of horizontal merging, the quantized DCT coefficient value of block 3 is copied to block 4 in step S210, and the DC coefficient and AC coefficient prediction process is performed in step S214.

If it is not horizontal merging, it is determined in step S211 whether the merger is a vertical merger. If it is not a vertical merging, the quantized DCT coefficient value of block 1 is copied to block 4 in step S212 and the process proceeds to step S214. In the case of vertical merging, the quantized DCT coefficient value of block 2 is copied to block 4 in step S213, and the process proceeds to step S214.

A prediction process of the DC coefficient and the AC coefficient is performed in step S214. If it is determined in step S215 that the last macroblock containing the video information is input, the process returns to step S201 to perform the following process. If not, the program is terminated.

The prediction process of the DC coefficient and the AC coefficient performed in step S214 of the flow chart will be described again for each block.

First, the merging information output from the object boundary block merging unit 111 (the object boundary block merging unit 152 in the case of the video signal encoding unit 144 of the VOP encoder) constituting the video signal encoding unit 84 of the encoder The block 2 is merely merged into the block 1, so that the quantized DCT coefficient values of the block 1 are copied to the block 2 for AC / DC predictive coding (COPY (S202-S203) and is used for the prediction process (S214).

Secondly, when the information that the block 3 has changed from the object boundary block to the object outside block is inputted (S204) by the merging information outputted from the object merge block merging units 111 and 152, the block 3 finds the merged block in the following order The quantized DCT coefficient values of the block are copied and used for the prediction process.

1) merging information is checked to see if it merged into block 1 by vertical merging (S205). If merged with block 1, the quantized DCT coefficient values of block 1 are copied to block 3 (S207).

2) If merging information has not been merged, the merged information is merged into block 2 in the diagonal direction. Therefore, the quantized DCT coefficient values of block 2 are copied to block 3 (S206).

Thirdly, if the merged information indicates that block 4 has changed from an object boundary block to an out-of-object block, block 4 finds the merged block in the following order, copies the quantized DCT coefficient values of the block, (S209-S213).

The merged information is firstly checked to see if the merger has been merged into the block 3 by the horizontal merger, and if the merger has been merged with the block 3, the quantized DCT coefficient values of the block 3 are copied to the block 4 (S210).

2) If merging information has not been merged, it is checked whether the merger is merged into the block 2 by the vertical merging (S211). If merging is performed with the block 2, the block 4 is merged with the quantized DCT The coefficient values are copied (S213).

3) If merging information has not been merged as in steps 1) and 2) above, the quantized DCT coefficient values of block 1 are copied to block 4 (S212) because they are merged into block 1 in a diagonal direction.

In another embodiment, when the transform coefficients and the encoding pattern coding units 114 and 155 generate texture information, the merging information output from the object boundary block merging units 111 and 152 may be merged into the object boundary block , The DC coefficient of the quantized DCT coefficient of the merged block is copied to the current sub-block and the AC coefficient is set to zero.

18, the steps S181 to S189 are the same. In the prediction process of step S190, only the DC coefficient of the quantized DCT coefficients of the merged block is copied to the current sub-block, and the AC coefficient is 0 is set.

In the prediction process of step S214, only the DC coefficient of the quantized DCT coefficient of the merged block is copied to the current sub-block, and the AC coefficient is set to 0 Setting.

As another embodiment, the DC coefficient is set to 2 ^{(bits_per_pixel-1)} and the AC coefficient is set to 0 without DCT coding like the original out-of-object block. Here, bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel.

That is, when an object boundary block merging process is applied and an object boundary block is transformed into an out-of-object block, the block is not DCT-encoded like an existing out-of-object block, but the DC component is 2 (bits_per_pixel -1), and the AC component is set to zero. 18 and 19 are applied, and steps S181 to S189 and steps S201 to S1213 are the same, and only the prediction processes of steps S190 and S214 are performed as described above.

In another embodiment, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-of-object block by the merging process, If there is a block that is not an out-of-block, the AC coefficient and the DC coefficient of the block are copied to the current sub-block. 18 and 19 are applied, and steps S181 to S189 and steps S201 to S1213 are the same, and only the prediction processes of steps S190 and S214 are performed as described above. The decoding method is also performed in the same manner as the flow chart shown in Figs. 18 and 19, and the block changed from the object boundary block to the out-object block is checked by the merging information extracted by the merging information extracting unit 175, And decodes it by performing the copying process described in FIG.

20 and 21 are flowcharts showing an embodiment of a decoding method for decoding and decoding after merging prediction according to the present invention.

As described above, according to the present invention, when an object boundary block merging technique is applied to an image information encoding / decoding unit, a block that is an out-of-object block in an object boundary block is efficiently processed by merging to improve coding / decoding efficiency Can be improved.

1 (A) - (F) show a hierarchical structure of a video signal.

FIG. 2 is an embodiment of an image having an arbitrary shape reconstructed as a macro block (16 × 16 pixels) for image encoding.

FIG. 3 shows a scanning method before applying the object boundary block merging technique.

4 is an embodiment showing a method of predicting the DC coefficient.

FIG. 5 is an embodiment showing a method of predicting the AC coefficient.

6 shows the types of sub-blocks in the macroblock.

FIG. 7A is a sub-block before the object boundary block merging technique is applied.

(B) shows that the subblock 4 changes from the object boundary block to the out-object block through the process of merging the object boundary block.

8 is a block diagram showing a configuration of an embodiment of an image information encoder according to the present invention.

FIG. 9A is a block diagram showing a configuration of an embodiment of a video signal encoding unit according to the present invention.

(B) is a block diagram showing the configuration of another embodiment of the video signal encoding unit according to the present invention.

10 is a scanning method for predicting the DC coefficient and the AC coefficient after applying the object boundary block merging technique.

11 is a block diagram showing a configuration of another embodiment of a video signal encoding unit according to the present invention for performing DC coefficient and AC coefficient prediction after merging.

12 is a block diagram showing a configuration of a video information decoder according to the present invention.

13 is a block diagram showing a configuration of a video signal reproducing unit of a decoder for decoding image information encoded by a merging and predicting process.

14 is a block diagram showing a configuration of a VOP encoder according to the present invention.

15 is a block diagram showing a configuration of another embodiment of a video signal encoding unit according to the present invention for performing DC coefficient and AC coefficient prediction after merging.

16 is a block diagram showing a configuration of a VOP decoder according to the present invention.

17 is a block diagram showing a configuration of a video signal decoding unit of a VOP decoder for decoding image information encoded by a merging and predicting process.

FIG. 18 is a flowchart showing an embodiment of a coding method for performing a post-merging prediction process according to the present invention.

FIG. 19 is a flowchart showing another embodiment of a coding method for performing a post-merging prediction process according to the present invention.

FIG. 20 is a flowchart illustrating a decoding method for decoding and decoding a merging prediction process according to an embodiment of the present invention.

FIG. 21 is a flowchart showing another embodiment of a decoding method for decoding and decoding after merging prediction according to the present invention.

Description of the Related Art

151 .... object boundary block padding 152 .... object boundary block merger

153 ... DCT transform unit 154 .... Quantization unit

155 .... Transform coefficient encoding unit 156 .... Inverse quantization unit

157 .... Inverse discrete cosine transform unit 158 .... Object boundary block separator

Claims (49)

A video information encoding apparatus for encoding an image, comprising: A motion estimator for receiving an object image signal and estimating motion in macroblock units; A motion compensation unit for compensating for motion of motion information of a macroblock estimated by the motion estimation unit; A subtractor for detecting a difference between the video signal of the macro block unit and the object video signal of the macro block unit, An image signal encoding unit for receiving a difference value detected by the subtractor and comparing the values of pixels corresponding to the blocks to determine whether the block is an object boundary block and performing an object boundary block merging process; An adder for adding a video signal of a macroblock unit compensated for motion in the motion compensation unit and internal information of an object encoded by the video signal encoding unit; A previous video signal detector for detecting a previous video signal and outputting the previous video signal to the motion estimator and the motion compensator, A multiplexer for multiplexing the motion information estimated by the motion estimator and the internal information of the object encoded by the image signal encoder, And a buffer for transmitting data output from the multiplexer. The apparatus of claim 1, wherein the image signal encoding unit An object boundary block merging unit which receives a video signal and compares values of corresponding pixels between the respective blocks to determine whether the object boundary block is an object boundary block, and performs an object boundary block merging process; A discrete cosine transform unit for receiving a video signal and performing discrete cosine transform; A quantization unit for receiving the quantized transform coefficients output from the DCT unit and quantizing the transform coefficients; The quantized transform coefficient output from the quantization unit and the merging information output from the object boundary block merging unit are received and compared with the values of pixels corresponding to each block to determine whether the object boundary block is an object boundary block to generate and output texture information A transform coefficient and an encoded block pattern coding unit; An inverse quantization unit for receiving the transform coefficients quantized by the quantization unit, dequantizing the transform coefficients, and extracting and outputting the transform coefficients; An inverse discrete cosine transform unit for performing inverse discrete cosine transform on the transform coefficients output from the inverse quantization unit; And an object boundary block separator for receiving the image signal output from the inverse discrete cosine transformation unit and the merged information in the object boundary block merging unit and outputting decoded data. The apparatus of claim 1, wherein the image signal encoding unit A macroblock scanning unit that scans a macroblock in units of subblocks and scans image information; A merging process performing unit in a macroblock merging when the scanned sub-block satisfies the merge condition; And an image coding unit for coding the merged block in the merging process performing unit in the macroblock. The apparatus of claim 1, wherein the image signal encoding unit A total image merging unit for performing a merging process in the entire image; And an image coding unit for coding the merged and outputted image signal in the whole image merging unit. The merging process is performed over the entire area without being restricted to the macro blocks in the independent encoding mode of the motion information and the image information, And outputs the encoded video information. 4. The method of claim 3, wherein the macroblock scan unit divides the macroblock into blocks 1 to 4 and stores the outline information of the object while scanning each sub-block, And performs the merging process by comparing the outline information of the corresponding coordinates between the boundary blocks. 4. The method according to claim 3, characterized in that if there is no overlapping pixel between any two object boundary blocks and one block is rotated 180 degrees from one block, then the former block is merged into the latter block and encoded Information encoding apparatus. A video information decoding apparatus for decoding encoded video information, A demultiplexer for multiplexing and separating multiplexed video signal information and motion information input from a coding unit; A motion decoding unit for receiving motion information output from the demultiplexer and detecting motion information; An image signal decoding unit which receives the image signal information output from the demultiplexer, compares values of pixels corresponding to the blocks, determines whether the block is an object boundary block, performs an object boundary block dividing process, and performs a prediction process to decode the object boundary block; A motion compensation unit for receiving motion information and image data detected by the motion decoding unit and compensating for motion; A motion reconstruction unit that receives motion compensation information output from the motion compensation unit and reproduces and outputs an image; And an image storage unit for storing the reproduced image from the image reproduction unit and outputting the stored image data to the motion compensation unit. 8. The apparatus of claim 7, wherein the image signal decoding unit The demultiplexer receives the input image information, compares the values of corresponding pixels between the blocks, determines whether the block is an object boundary block, performs an object boundary block dividing process, and performs a prediction process to decode the image information to output a transform coefficient And an encoding pattern decoding unit; An inverse quantization unit for inversely quantizing the transform coefficients and the image information decoded and output by the encoding pattern decoding unit by multiplying them by a quantization coefficient; An inverse quantization unit that inversely transforms the inversely quantized transform coefficients output from the inverse quantization unit and outputs the transformed inverse transformed image data as image data in a spatial domain; And an object boundary block separator for receiving the video signal output from the IDCT unit and outputting the decoded video data in units of macroblocks. 1. An image information encoding apparatus for separating and encoding an image into an object image and a background image having arbitrary shape information, A VOP forming unit for receiving a video signal to form a VOP; a motion estimator for estimating a motion of a macroblock unit output from the VOP forming unit; A motion compensator for receiving and motion compensating the motion information output from the motion estimator; A subtractor for subtracting a motion compensated VOP output from the motion compensation unit and a VOP formed and output from the VOP forming unit, An image signal encoding unit for receiving a difference value and shape information output from the subtractor, applying an object boundary merging technique to a macroblock to perform a merging process, and encoding the merged macroblock in units of subblocks, An adder for adding a motion compensated VOP in the motion compensation unit and internal information of an object encoded by the video signal encoding unit; A previous VOP detection unit for detecting a VOP of a previous screen by receiving an output signal of the adder and outputting a VOP of the detected previous screen to the motion estimation unit and the motion compensation unit, A shape information encoding unit that encodes shape information for a macroblock-based image output from the VOP forming unit; A multiplexer for multiplexing and outputting motion information estimated by the motion estimator, internal information of an object hatched by the image signal encoder and shape information encoded by the shape information protector, And a buffer for transmitting information output from the multiplexer in a bitstream. The apparatus of claim 9, wherein the output signal of the shape information encoding unit is input to the motion estimation unit, the motion compensation unit, and the video signal encoding unit to perform motion estimation, motion compensation, Information encoding apparatus. The apparatus of claim 9, wherein the image signal encoding unit A macroblock scanning unit that scans a macroblock in units of subblocks and scans image information; A merging process performing unit in a macroblock merging when the scanned sub-block satisfies the merge condition; And an image coding unit for coding the merged block in the merging process performing unit in the macroblock. The apparatus of claim 9, wherein the image signal encoding unit comprises a full image merging unit for performing a merging process in an entire image, and an image-based image information encoding unit for encoding a merged and outputted image signal in the full image merging unit, Wherein encoding is performed by performing a merging process over the entire area of the VOP without being limited to macro blocks as in the independent encoding mode of information and image information. The apparatus of claim 9, wherein the image signal encoding unit A boundary block padding unit for receiving a video signal and original shape information and setting a non-object part in the sub-block as an average value of the object part, A boundary block merging unit that receives the image signal output from the boundary block padding unit and the reproduced shape information and performs an object boundary block merging process, A discrete cosine transformation unit that receives the image signal output from the object boundary block merging unit and performs discrete cosine transformation, A quantization unit for receiving and quantizing the transform coefficients output from the DCT unit, A transform coefficient and an encoding pattern encoding unit for encoding quantized transform coefficients output from the quantization unit and merging information output from the object boundary block merging unit and encoding the texture information, An inverse quantization unit for receiving the transform coefficients quantized by the quantization unit and dequantizing the transform coefficients, An inverse quantization unit for performing inverse discrete cosine transform on the transform coefficients output from the inverse quantization unit, And an object boundary block separator for receiving the image signal output from the inverse discrete cosine transformation unit and the reconstructed shape information and outputting decoded data. 14. The method of claim 13, wherein, when the transform coefficient and the encoding pattern encoding unit generate the texture information, the merging information output from the object boundary block merging unit is merged into the object boundary block in the object boundary block (DC and AC coefficients) of the merged block to the current sub-block, when it is indicated that the block is a changed block. 14. The method of claim 13, wherein, when the transform coefficient and the encoding pattern encoding unit generate texture information, when the merging information indicates that the current sub-block is a block changed from an object boundary block to an out- The DC coefficient is set to 2 ^{(bits_per_pixel-1)} (where bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel) and the AC coefficient is set to 0 And outputs the image information. 14. The method according to claim 13, wherein, when the transform coefficient and the encoding pattern encoding unit generate texture information, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the object outside block by the merging process, If there is a block other than an out-of-object block among the left block and the upper block without DCT coding as in the outer block, the AC coefficient and the DC coefficient of the block are copied to the current sub-block and the prediction process is performed Device. 14. The method according to claim 13, wherein when the transform coefficient and the encoding pattern encoding unit generate texture information, if the merging information indicates that the current sub-block is a block changed from the object boundary block to the object outside block by the merging process, Wherein the predictive process is performed by copying the quantized DC coefficient of the current block into the current sub-block and setting the AC coefficient to 0. 14. The method according to claim 13, wherein when the transform coefficient and the encoding pattern encoding unit generate texture information, if the merging information indicates that the current sub-block is a block changed from the object boundary block to the object outside block by the merging process, And a prediction process is performed by copying the quantized DCT coefficients of the current block to the current sub-block. 1. An image information decoding apparatus for decoding image information separately coded into an object image having arbitrary shape information and a background image, A demultiplexer for multiplexing and demultiplexing the transmitted multiplexed image information; A shape information decoding unit for decoding and outputting the image information output from the demultiplexer; A motion decoding unit for detecting motion information from the image information output from the demultiplexer; An image signal decoding unit decoding the image signal from the image information output from the demultiplexer and outputting the decoded image signal; A motion compensation unit for receiving motion information detected by the motion decoding unit, shape information output from the shape information decoding unit and VOP data, and compensating for motion; A VOP reconstructing unit for reconstructing a VOP by separating blocks merged with the shape information output from the shape information decoding unit, the motion compensation information output from the motion compensation unit, and the video signal output from the video signal decoding unit, And a decoding unit for decoding the image information. 20. The apparatus of claim 19, wherein the VOP representation comprises: A merging information extracting unit that receives shape information output from the shape information decoding unit and extracts merging information by the same merging process as the encoding unit and outputs the merging information; A merging information extracting unit for merging information, a motion compensation information output from the motion compensating unit, and a transform coefficient and a coding pattern decoding unit for decoding and outputting the encoded video information received from the video signal decoding unit, ; An inverse quantization unit for inverse-quantizing the transform coefficients and the image information decoded by the encoding pattern decoder by a quantization coefficient; An inverse quantization unit that inversely transforms the inversely quantized transform coefficients output from the inverse quantization unit and outputs the transformed inverse transformed image data as image data in a spatial domain; And an object boundary block separator for receiving the decoded video signal output from the IDCT and the reproduced shape information output from the shape information decoding unit and outputting the decoded VOP data in units of macroblocks. Information decoding apparatus. 21. The method of claim 20, wherein, when generating the texture information, the conversion coefficient and the encoding pattern decoding unit display, when the merging information indicates that the current sub-block is a block changed from an object boundary block to an out- The DC coefficient is set to 2 ^{(bits_per_pixel-1)} (where bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel) and the AC coefficient is set to 0 And decodes the video information. 21. The method according to claim 20, wherein when the texture information is generated, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the object outside block by the merging process, If there is a block other than an out-of-object block in the left block and the upper block, the AC coefficient and the DC coefficient of the block are copied to the current sub-block and a prediction process is performed to decode the same Information decoding apparatus. 21. The method as claimed in claim 20, wherein, in generating texture information, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the object outside block by the merging process, And the quantized DC coefficient of the current block is copied to the current sub-block, and the AC coefficient is set to 0, and a prediction process is performed to decode the decoded DC coefficient. 21. The method as claimed in claim 20, wherein, in generating texture information, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the object outside block by the merging process, Block to a current sub-block and performing a prediction process to decode the quantized DCT coefficient. 1. An image information encoding method for separating an image into an object image and a background image having arbitrary shape information, When the texture information is generated, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the outside object block by the merging process, The DC coefficient is set to 2 ^{(bits_per_pixel-1)} (where bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel) and the AC coefficient is set to 0 And setting the image information. 1. An image information encoding method for separating an image into an object image and a background image having arbitrary shape information, When texture information is generated, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-of-object block by the merging process, And if there is a block other than an out-of-object block, the AC coefficient and the DC coefficient of the block are copied to the current sub-block to perform the prediction process. 1. An image information encoding method for separating an image into an object image and a background image having arbitrary shape information, When the texture information is generated, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-of-object block by the merging process, copy the quantized DC coefficient of the merged block to the current sub- And an AC coefficient is set to 0 to perform a prediction process. 1. An image information encoding method for separating an image into an object image and a background image having arbitrary shape information, When the texture information is generated, when the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-of-object block by the merging process, copy the quantized DCT coefficients of the merged block to the current sub- And performing a prediction process on the image information. 28. The method of claim 25, wherein when the object boundary block merging information indicates that the block 2 has changed into an out-of-object block by the object boundary block merging process, the quantized DCT coefficient values of the block 1 merged in the block 2 Copying and performing the prediction process, In the case of indicating that block 3 has changed into an out-of-object block by the object boundary block merging process, the block merged in the following order is found, the quantized DCT coefficient values of the block are copied, 1) Object boundary block merging information is used to check whether merged into block 1 by vertical merger, and then block 3 is copied with quantized DCT coefficient values of block 1. 2) If it is not merged as described in 1) above using the object boundary block merging information, since it has been merged into the block 2 in the diagonal direction, the quantized DCT coefficient values of the block 2 are copied to the block 3 for AC / DC predictive coding . If it is indicated that the block 4 has changed into an out-of-object block by the object boundary block merging process, the block merged in the following order is found, the quantized DCT coefficient values of the block are copied, First, we check whether the block merged in block 3 by merging horizontally using object boundary block merging information. If it is merged with block 3, copy the quantized DCT coefficients in block 3 into block 4. 2) If it is not merged as in the process 1) above using the merging information of the object boundary block, it is checked whether the merger is performed in the block 2 by the vertical merger, and if the merger is performed with the block 2, the quantized DCT coefficient of the block 2 Copies the values. 3) If it is not merged as in 1) and 2) above using the object boundary block merging information, since it is merged into block 1 in the diagonal direction, the quantized DCT coefficient values of block 1 are copied to block 4, And performing an encoding process on the image information. A video information encoding method for encoding an image, When the merging information output from the object boundary block merging unit is received and it is determined from the merging information that the block 2 has changed from the object boundary block to the out-of-object block, the quantized DCT coefficient value of the block 1 is converted into the DCT coefficient of the block 2 ; ≪ / RTI > If the block 2 has not changed from the object boundary block to the out-of-object block and the block 3 has changed from the merging information to the out-of-object block, it is determined whether the merging is vertical merging. If the merging is vertical merging, Copying the coefficient value to block 3; Copying the quantized coefficient value of block 2 to block 3 if it is not a vertical merging; If it is determined that block 3 has not changed from the merging information to the out-of-object block, it is determined whether block 4 has changed from the object boundary block to the out-of-object block. If block 4 is changed from the object boundary block to the out- And copying the quantized DCT coefficient value of the block 3 to the block 4 if the horizontal merging is judged; Determining whether the merger is a vertical merge in the case where the merger is not a horizontal merge and copying the quantized DCT coefficient value of the block 1 to the block 4 and performing a DC coefficient and AC coefficient predicting process; In the case of vertical merging in the step, copying the quantized DCT coefficient values of block 2 to block 4 and performing a DC coefficient and AC coefficient prediction process; The DC coefficient and the AC coefficient are predicted, and it is determined whether the next macro block image information is the last macro block including the image information. If the macro block image information is not the last macro block, And returning to the step of receiving the information and performing the following process, and terminating the present program if the last macroblock is present. The method of claim 30, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merge process, The DC coefficient is set to 2 ^{(bits_per_pixel-1)} (where bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel) and the AC coefficient is set to 0 And a prediction process is performed by setting the motion vector. The method of claim 30, wherein, when the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by the merging process, If there is a block other than an out-of-object block among the left block and the upper block without DCT coding, the AC coefficient and the DC coefficient of the block are copied to the current sub-block to perform a prediction process. The method of claim 30, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-matter block by the merging process, The DC coefficient is copied to the current sub-block, and the AC coefficient is set to 0 to perform the prediction process. The method of claim 30, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by the merging process, And performing a prediction process by copying the DCT coefficients to the current sub-block. A video information encoding method for encoding an image, When merging information output from the object boundary block merging unit is received and it is determined whether block 1 has changed from an object boundary block to an out-of-object block from the merging information, and the quantized DCT coefficients Value to the block 1 and proceeding to the next step, and if not, proceed directly to the next step; If it is determined that the block 2 has changed from the merging information to the out-of-object block, the quantized DCT coefficient value of the block in which the information of the block 2 is merged is copied to the block 2 and the process proceeds to the next step. , Directly proceeding to the next step; If it is determined that the block 3 has changed from the merging information to the out-of-object block, the quantized DCT coefficient value of the block in which the information of the block 3 is merged is copied to the block 3, and proceeds to the next step. , Directly proceeding to the next step; If it is determined that the block 4 has changed from the merging information to the out-of-object block, the quantized DCT coefficient value of the block in which the information of the block 4 is merged is copied to the block 4, and the process proceeds to the next step. , Directly proceeding to the next step; DC coefficient and an AC coefficient prediction process of the image information. The method of claim 35, wherein, when the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merge process, The DC coefficient is set to 2 ^{(bits_per_pixel-1)} (where bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel) and the AC coefficient is set to 0 And setting the image information. The method as claimed in claim 35, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by the merging process, If there is a block other than an out-of-object block among the left block and the upper block without performing DCT coding, the AC coefficient and the DC coefficient of the block are copied to the current sub-block and the prediction process is performed. The method of claim 35, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merging process, Wherein the prediction process is performed by copying the quantized DC coefficient to the current sub-block and setting the AC coefficient to zero. The method of claim 35, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merging process, And the quantized DCT coefficients are copied to the current sub-block to perform a prediction process. In the video information decoding method, If the merging information is received and the block 1 is changed from the object boundary block to the out-of-object block from the merging information, the quantized DCT coefficient value of the block in which the information of the block 1 is merged is copied to the block 1 Proceeding to the next step and proceeding directly to the next step if not changed in the determination; If it is determined that the block 2 has changed from the merging information to the out-of-object block, the quantized DCT coefficient value of the block in which the information of the block 2 is merged is copied to the block 2 and the process proceeds to the next step. , Directly proceeding to the next step; If it is determined that the block 3 has changed from the merging information to the out-of-object block, the quantized DCT coefficient value of the block in which the information of the block 3 is merged is copied to the block 3, and proceeds to the next step. , Directly proceeding to the next step; If it is determined that the block 4 has changed from the merging information to the out-of-object block, the quantized DCT coefficient value of the block in which the information of the block 4 is merged is copied to the block 4, and the process proceeds to the next step. , Directly proceeding to the next step; DC coefficient, and AC coefficient prediction of the image data. The method of claim 40, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by the merging process, The DC coefficient is set to 2 ^{(bits_per_pixel-1)} (where bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel) and the AC coefficient is set to 0 And a prediction process is performed by setting the motion vector. The method of claim 40, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merge process, If there is a block other than an out-of-object block in the left block and the upper block without performing DCT coding, the AC coefficient and the DC coefficient of the block are copied to the current sub-block and the prediction process is performed. The method of claim 40, wherein, in the case where the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-object block by the merging process, A DC coefficient is copied to a current sub-block, and an AC coefficient is set to 0 to perform a prediction process. The method of claim 40, wherein, in the case where the merging information indicates that the current sub-block is a block changed from the object boundary block to the out-object block by the merging process, And performing a prediction process by copying the generated DCT coefficients to the current sub-block. In the video information decoding method, Copying the quantized DCT coefficient value of the block 1 to the DCT coefficient of the block 2 when the merging information is received and the block 2 is changed from the merging information to an out-of-object block in the object boundary block; If the block 2 has not changed from the object boundary block to the out-of-object block and the block 3 has changed from the merging information to the out-of-object block, it is determined whether the merging is vertical merging. If the merging is vertical merging, Copying the coefficient value to block 3; Copying the quantized coefficient value of block 2 to block 3 if it is not a vertical merging; If it is determined that block 3 has not changed from the merging information to the out-of-object block, it is determined whether block 4 has changed from the object boundary block to the out-of-object block. If block 4 is changed from the object boundary block to the out- And copying the quantized DCT coefficient value of the block 3 to the block 4 if the horizontal merging is judged; Determining whether the merger is a vertical merge in the case where the merger is not a horizontal merge and copying the quantized DCT coefficient value of the block 1 to the block 4 and performing a DC coefficient and AC coefficient predicting process; In the case of vertical merging in the step, copying the quantized DCT coefficient values of block 2 to block 4 and performing a DC coefficient and AC coefficient prediction process; The DC coefficient and the AC coefficient are predicted, and it is determined whether the next macro block image information is the last macro block including the image information. If the macro block image information is not the last macro block, And returning to the step of receiving information and performing the following process, and terminating the present program if the last macroblock is present. 46. The method of claim 45, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-matter block by a merge process, The DC coefficient is set to 2 ^{(bits_per_pixel-1)} (where bits_per_pixel is the number of bits representing PCM (Pulse Code Modulation) data of one pixel) and the AC coefficient is set to 0 And decodes the decoded video information. The method as claimed in claim 45, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merge process, If there is a block other than an out-of-object block among the left block and the upper block without performing DCT coding, the AC coefficient and the DC coefficient of the block are copied to the current sub-block and the prediction process is performed. The method of claim 45, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merging process, Wherein the prediction process is performed by copying the quantized DC coefficient to the current sub-block and setting the AC coefficient to zero. The method of claim 45, wherein, in the case where the merging information indicates that the current sub-block is a block changed from an object boundary block to an out-of-object block by a merging process, And a prediction process is performed by copying the quantized DCT coefficients to the current sub-block.

Images