KR100739757B1 - Image coding/decoding method and apparatus thereof - Google Patents

Abstract

The present invention relates to a video encoding and decoding method and apparatus capable of improving the compression rate for video data by reducing the number of bits of a motion vector allocated to a bitstream in units of slices. If received, detecting a flag indicating a motion vector allocation unit from the received bitstream; Unpacking the received bitstream on a slice basis or on a macroblock basis based on the detected flag and inverse quantizing; And compensating for the motion of the dequantized bitstream using the motion vector included in the dequantized bitstream, thereby improving the compression ratio of the image data by reducing the number of bits for the motion vector included in the bitstream. have.

Description

Image coding and decoding method and apparatus thereof

1 shows an example of dividing a slice into 16 macroblocks.

2 is an example of a conventional bitstream structure for one slice.

3 is a functional block diagram of an image decoding apparatus according to an embodiment of the present invention.

4 (a) and 4 (b) show examples of a compressed bitstream structure input to the video decoding apparatus shown in FIG.

FIG. 5 is a functional block diagram of an image encoding apparatus corresponding to the image decoding apparatus illustrated in FIG. 3.

6 is an exemplary diagram of grouping of macroblocks in one slice.

7 is an example of a bitstream structure when a group of macroblocks is formed in a slice.

FIG. 8 is a functional block diagram of an image encoding apparatus corresponding to a case in which the image decoding apparatus illustrated in FIG. 3 is properly implemented to decode the bitstream illustrated in FIG. 7.

9 is a functional block diagram of an image decoding apparatus according to another embodiment of the present invention.

10 (a) and 10 (b) show examples of a bitstream structure in which a flag indicating a motion vector allocation unit is allocated in a slice unit and a macroblock unit.

FIG. 11 is a functional block diagram of an image encoding apparatus corresponding to the image decoding apparatus illustrated in FIG. 9.

12 is a flowchart illustrating an image decoding method according to another embodiment of the present invention.

FIG. 13 illustrates an example in which the image decoding method illustrated in FIG. 12 is applied when a flag indicating a motion vector allocation unit is detected in a slice unit.

FIG. 14 illustrates an example in which the image decoding method illustrated in FIG. 12 is applied when a flag indicating a motion vector allocation unit is detected in macroblock units.

FIG. 15 illustrates an example in which the image decoding method illustrated in FIG. 12 is applied when a flag indicating a motion vector allocation unit is detected in a slice unit and a macroblock unit.

The present invention relates to a method and apparatus for image encoding and decoding, and more particularly, to motion detection by motion estimation during image encoding, and motion compensation by using motion vector during image decoding. The present invention relates to a video encoding and decoding method and apparatus therefor.

The conventional video encoding method according to the standards such as MPEG-1 / 2/4, H.261, H.263, and H.264 divides one slice into several macro blocks as shown in FIG. 1 illustrates an example of dividing a slice into 16 macro blocks MB0 to MB15. The image encoding method detects an independent motion vector by performing motion estimation on each of the divided macroblocks through Rate Distortion (RD) optimization. Accordingly, as shown in FIG. 2, the bitstream of one slice has a structure of allocating a number of motion vectors corresponding to the number of macroblocks. 2 is an example of a conventional bitstream structure when one slice is divided into 16 macroblocks.

In addition, as shown in FIG. 1, the conventional video encoding method may divide each macro block into sub-blocks of a smaller unit through the RD optimization. If each macroblock is divided into smaller subblocks, the conventional video encoding method detects one motion vector independent for each subblock by motion estimation, so that a bitstream of one slice is sliced. It has a structure for allocating a number of motion vectors corresponding to the number of sub-blocks included in.

However, as mentioned above, the allocation of the motion vector becomes a factor that makes it difficult to improve the compression ratio of the image data.

An object of the present invention is to provide an image encoding and decoding method and apparatus capable of improving the compression ratio for image data by reducing the number of bits of a motion vector allocated to a bitstream in slice units.

According to an aspect of the present invention, there is provided a method of decoding an image, the method comprising: detecting a flag indicating a motion vector allocation unit from the received bitstream when the bitstream is received; Unpacking and inverse quantizing the received bitstream in units of slices or macroblocks based on the detected flag; And compensating for the motion of the dequantized bitstream using the motion vector included in the dequantized bitstream.

According to an aspect of the present invention, in the image decoding apparatus, when a bitstream is received, the received bitstream is sliced or based on a flag indicating a motion vector allocation unit included in the received bitstream. A bitstream unpacking unit which unpacks the macroblock unit; An inverse quantizer for inversely quantizing the unpacked bitstream in units of slices or units of macroblocks; And a motion compensator for compensating for motion of the dequantized bitstream in the dequantizer, wherein a flag indicating a motion vector allocation unit is allocated in a slice unit and / or a macroblock unit, and the motion compensation unit is in a slice unit. Compensate motion for all macroblocks included in the slice using dequantized motion vectors, or compensate for motion for macroblocks using dequantized motion vectors in macroblock units, or dequantize groups in macroblock units. The present invention provides an image decoding apparatus using a motion vector to compensate for a motion of a macroblock included in a group of macroblocks.

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

3 is a functional block diagram of a motion decoding apparatus according to an embodiment of the present invention. Referring to FIG. 3, the motion decoding apparatus includes an unpacking unit 310, a de-quantization unit 320, and a motion compensator 330.

In the motion decoding apparatus illustrated in FIG. 3, a flag indicating a motion vector allocation unit is allocated in a slice unit. Therefore, the bitstream input to the motion decoding apparatus shown in FIG. 3 may be configured as shown in FIG. 4 (a) or 4 (b).

4 (a) shows an example of a bitstream structure in which a flag indicating a motion vector allocation unit is allocated in units of slices, and a motion vector is allocated in macroblock units. FIG. An example of a bitstream structure in which units are allocated in units and one motion vector is allocated to one slice. A flag (SMFlag, Slice Motion Flag) indicating a motion vector allocation unit allocated in a slice unit is included in the slice data area as shown in FIGS. 4 (a) to 4 (b). The flag SMFlag may be defined as 1 bit.

When a bitstream is received, the unpacking unit 310 divides the received bitstream into slices or macroblocks based on a flag indicating a motion vector allocation unit included in the received bitstream. Unpack with.

To this end, the unpacking unit 310 may include a flag detector 311, a first unpacking processor 312, and a second unpacking processor 313 as shown in FIG. 3.

When the bitstream is received, the flag detector 311 detects a flag indicating a motion vector allocation unit in the received bitstream. If the flag indicating the detected motion vector allocation unit is a slice unit flag SMFlag and the flag SMFlag value is "0", the flag detector 311 transmits the received bitstream to the first unpacking processor 312. do. If the flag SMFlag indicating the motion vector allocation unit is "0", the flag indicating the motion vector allocation unit indicates that the motion vector is allocated in macroblock units. When the flag SMFlag indicating the motion vector allocation unit is "0", the received bitstream has a structure as shown in Fig. 4A.

If the flag indicating the detected motion vector allocation unit is a slice unit flag SMFlag, and the value of the flag SMFlag is "1", the flag detector 311 performs a second unpacking processor 313 on the received bitstream. To send. If the flag SMFlag indicating the motion vector allocation unit is "1", the flag SMFlag indicating the motion vector allocation unit indicates that a motion vector is allocated one motion vector per slice. When the flag SMFlag indicating the motion vector allocation unit is "1", the received bitstream has a structure as shown in Fig. 4B.

The first unpacking processor 312 unpacks and outputs the bitstream in macroblock units. The output unpacked bitstream is transmitted to the first dequantizer 321 of the dequantizer 320.

The second unpacking processor 313 unpacks and outputs the bitstream in slice units. The output unpacked bitstream is transmitted to the second dequantizer 322 of the dequantizer 320.

The dequantizer 320 dequantizes the unpacked bitstream in units of slices or macroblocks. To this end, the inverse quantization unit 320 includes a first inverse quantizer 321 and a second inverse quantizer 322 as shown in FIG. 3.

The first dequantizer 321 dequantizes the unpacked bitstream transmitted from the first unpacking processor 312 in macroblock units. Accordingly, the motion vector allocated in units of macro blocks is also dequantized. The dequantized bitstream is output to the motion compensator 330.

The second dequantizer 322 dequantizes the unpacked bitstream transmitted from the second unpacking processor 312 in slice units. Accordingly, the motion vector allocated in the slice unit is also dequantized. The dequantized bitstream is output to the motion compensator 330.

When the dequantized bitstream is received from the first inverse quantizer 321, the motion compensator 330 performs motion compensation on the macroblock by using the dequantized motion vector in macroblock units and performs motion compensation. Output the video signal. When the inverse quantized bitstream is received from the second inverse quantizer 322, the motion compensator 330 performs motion compensation on the macro block by using the inversely quantized motion vector in units of slices, and the motion compensated image. Output the signal.

FIG. 5 is a functional block diagram of an image encoding apparatus corresponding to the image decoding apparatus illustrated in FIG. 3. That is, FIG. 5 is a functional block diagram of an image encoding apparatus for allocating a flag indicating a motion vector allocation unit in a slice unit and allocating a motion vector in a macroblock unit or a slice unit according to the allocated flag.

Referring to FIG. 5, the apparatus for encoding an image includes a slice-based motion estimator 501, a block-based motion estimator 502, a motion estimation type selection and flag value determiner 503, a quantizer 504, and a bit. A stream packing section 505.

The slice-based motion estimator 501 detects one motion vector per slice by performing motion estimation on a slice basis with respect to the input image signal. The motion estimated video signal is transmitted to the motion estimation type selection and flag value determination unit 503.

The block-based motion estimator 502 detects one motion vector per macro block by performing motion estimation on a macro block basis on an input image signal. The motion estimated video signal is transmitted to the motion estimation type selection and flag value determination unit 503.

The slice-based motion estimator 501 and the block-based motion estimator 502 output the RD cost obtained by rate-distortion (RD) optimization together with the motion estimation result.

The motion estimation type selection and flag value determination unit 503 compares the RD cost output from the slice-based motion estimation unit 501 with the RD cost output from the block-based motion estimation unit 502 and outputs a small value. Select the estimation type as the optimal type.

That is, if the RD cost output from the slice-based motion estimation unit 501 is smaller than the RD cost output from the block-based motion estimation unit 502, the motion estimation type selection and flag value determination unit 503 may slice-based motion estimation. The motion estimation type in the government unit 501 is selected, and the value of the slice unit flag SMFlag is determined as "1". Next, the motion estimation type selection and flag value determination unit 503 transmits the motion estimation result output from the slice-based motion estimation unit 501 to the quantization unit 504, and transmits the value of the flag SMFlag in the determined slice unit. The bitstream packing unit 505 is provided.

On the other hand, if the RD cost output from the block-based motion estimation unit 502 is smaller than the RD cost output from the slice-based motion estimation unit 501, the motion estimation type selection and flag value determination unit 503 may perform block-based motion. The motion estimation type in the estimation unit 502 is selected, and the slice unit flag (SMFlag) value is determined as "0". Next, the motion estimation type selection and flag value determination unit 503 transmits the determined flag value to the bitstream packing unit 505 while transmitting the motion estimation result output from the block-based motion estimation unit 502 to the quantization unit 504. to provide.

The quantization unit 504 quantizes and outputs motion estimation image data transmitted from the motion estimation type selection and flag value determination unit 503.

The bitstream packing unit 505 may perform a bitstream of the quantized image data output from the quantization unit 504 based on the flag value provided from the motion estimation type selection and the flag value determination unit 503. Packed and output as shown in 4 (b). That is, when the slice unit flag SMFlag value is "0", the bitstream packing unit 505 packs the bitstream of the quantized video signal as shown in FIG. Accordingly, the bitstream packing unit 505 outputs a bitstream to which a motion vector for each macroblock is allocated.

If the slice unit flag SMFlag value is "1", the bitstream packing unit 505 packs the bitstream of the quantized video signal as shown in FIG. Accordingly, the bitstream packing unit 505 outputs a bitstream in which one motion vector is allocated to one slice.

Meanwhile, the apparatus for decoding an image shown in FIG. 3 may also be applied when a flag GMBMFlag indicating a motion vector allocation unit is inserted in a macroblock unit. When a flag indicating a motion vector allocation unit is inserted into a bitstream in macroblock units, grouping is possible because the motion vectors of some macroblocks included in a slice are the same or similar.

6 is an exemplary diagram of grouping of macroblocks in one slice. 6 is a group of macroblocks in which one slice is divided into 16 macroblocks MB0 through MB15, and the macroblocks MB1, MB9, MB10, and MB14 of the divided macroblocks have the same or similar motion vectors. Is an example of forming.

When a group of macroblocks is formed as shown in FIG. 6, the bitstream input to the motion decoding apparatus shown in FIG. 3 has a structure as shown in FIG. 7. Referring to FIG. 7, a macroblock (GMBMFlag, Group of MacroBlock Moving Flag) indicating a motion vector allocation unit is set to “0” for a macroblock included in a slice among macroblocks included in one slice. Each motion vector is inserted. On the other hand, the macroblock included in the macroblock group among the macroblocks included in the slice has the flag (GMBMFlag) indicating the motion vector allocation unit set to "1", and the first macroblock among the macroblocks included in the macroblock group. In the block, a representative motion vector assigned to a group of macro blocks is inserted. The macro block unit flag GMBMFlag may be defined as 1 bit.

When the bitstream having the structure shown in FIG. 7 is received, the unpacking unit 310 of FIG. 3 macros the received bitstream based on a flag indicating a motion vector allocation unit included in the received bitstream. Unpack in units of blocks or groups of macro blocks.

To this end, when the bitstream is received, the flag detector 311 of the unpacking unit 310 detects a flag indicating a motion vector allocation unit in the received bitstream. At this time, if the flag indicating the detected motion vector allocation unit is a macroblock unit flag (GMBMFlag) and the flag value is "0", the flag detector 311 sends the received bitstream to the first unpacking processor 312. send. When the macroblock unit flag GMBMFlag is "0", it indicates that the corresponding macroblock is not included in the group of the macroblock.

If the flag indicating the detected motion vector allocation unit is a macro block unit flag (GMBMFlag) and the flag (GMBMFlag) value is "1", the flag detector 311 determines the received bitstream as a second unpacking processor 313. To send). When the macroblock unit flag GMBMFlag is "1", it indicates that a corresponding macroblock is included in a group of macroblocks.

The first unpacking processor 312 unpacks and outputs the bitstream in macroblock units. The output unpacked bitstream is transmitted to the first dequantizer 321 of the dequantizer 320.

The second unpacking processor 313 unpacks and outputs the bitstream in macroblock units. At this time, the second unpacking processor 313 considers that only the first macroblock among the macroblocks included in the group of macroblocks has a representative motion vector assigned to the group of macroblocks inserted into the bitstream. Unpack the. That is, the second unpacking processor 313 does not insert a motion vector in the macro blocks other than the first macro block among the macro blocks included in the group of macro blocks, and thus, the motion vector for the macro blocks other than the first macro block. Does not unpack. The output unpacked bitstream is transmitted to the second dequantizer 322 of the dequantizer 320.

The first dequantizer 321 dequantizes the unpacked bitstream transmitted from the first unpacking processor 312 in macroblock units. Accordingly, the motion vector is dequantized in macroblock units. The dequantized bitstream is output to the motion compensator 330.

The second dequantizer 322 dequantizes the unpacked bitstream transmitted from the second unpacking processor 312 in groups of macro blocks. Accordingly, the motion vector is inversely quantized in group units of the macro block. The dequantized bitstream is output to the motion compensator 330.

When the dequantized bitstream is received from the first inverse quantizer 321, the motion compensator 330 performs motion compensation on the macroblock by using the dequantized motion vector in macroblock units and performs motion compensation. Output the video signal. When the dequantized bitstream is received from the second inverse quantizer 322, the motion compensator 330 performs motion compensation on the macroblock by using the dequantized motion vector in group units of the macroblock, and the motion The compensated video signal is output.

8 is a functional block diagram of a video encoding apparatus corresponding to FIG. 3 when it is defined as an apparatus for decoding a bitstream in which a flag indicating a motion vector allocation unit is inserted in a macro block unit as described above.

Referring to FIG. 8, the apparatus for encoding an image includes a block-based motion estimator 801, a grouping processor 802 for a macroblock, a quantization unit 803, and a bitstream packing unit 804.

The block-based motion estimation unit 801 is configured and operates in the same manner as the block-based motion estimation unit 502 of FIG. 5. Accordingly, the block-based motion estimation unit 801 outputs a motion estimation result including a motion vector detected in units of macro blocks with respect to an input image signal.

The macroblock grouping processor 802 compares the motion vectors of each macroblock based on the result estimated by the block-based motion estimation unit 801, and groups the macroblocks in which the motion vectors are considered to be the same. Based on the grouping result, a flag (GMBMFlag) indicating a motion vector allocation unit is determined in macroblock units, and the motion estimation result transmitted from the block-based motion estimation unit 801 is provided while providing the determined flag to the bitstream packing unit 804. Transmit to quantizer 803.

The quantization unit 803 quantizes and outputs the motion estimation result in units of macroblocks. The output quantization result is transmitted to the bitstream packing unit 804.

The bitstream packing unit 804 packs the bitstream of the quantized value transmitted from the quantization unit 803 based on the flag value transmitted from the grouping processing unit 802 of the macro block. Accordingly, a bitstream having a structure as shown in FIG. 7 may be output.

9 is a functional block diagram of an image decoding apparatus according to another embodiment of the present invention. In the motion decoding apparatus illustrated in FIG. 9, a flag indicating a motion vector allocation unit is allocated in a slice unit and a macroblock unit.

Referring to FIG. 9, the motion decoding apparatus includes an unpacking unit 910, an inverse quantization unit 920, and a motion compensation unit 930.

The unpacking unit 910 unpacks the received bitstream in units of slices or macroblocks based on a flag indicating a motion vector allocation unit included in the received bitstream. Unpacking of macroblock units is performed by unpacking macroblocks included in a group of macroblocks and unpacking of macroblocks not included in a group of macroblocks. Accordingly, the unpacking unit 910 includes a flag detector 911, a first unpacking processor 912, a second unpacking processor 913, and a third unpacking processor 914 as shown in FIG. 9. do.

When the bitstream is received, the flag detector 911 detects a flag indicating a motion vector allocation unit in the received bitstream. In the case of FIG. 9, the received bitstream has a structure as shown in FIG. 10 (a) or FIG. 10 (b). FIG. 10A illustrates a bitstream structure in which a flag indicating a motion vector allocation unit is allocated in a slice unit when one motion vector is allocated to one slice. FIG. In a case where a group is defined, a flag indicating a motion vector allocation unit is a bitstream structure in which slices and macroblocks are allocated.

Accordingly, the flag detector 911 transmits the received bitstream to the first unpacking processor 912 when the detected flag is a flag SMFlag allocated in a slice unit and the flag SMFlag value is "1". .

The first unpacking processor 912 unpacks and receives the received bitstream in slice units. The output unpacked bitstream is transmitted to the first dequantizer 921 of the dequantizer 920.

The flag detector 911 detects a flag indicating a motion vector allocation unit again from the received bitstream when the detected flag is a flag SMFlag allocated in slice units and the flag SMFlag value is "0". Accordingly, the flag GMBMFlag allocated in macroblock units is detected, and if the detected flag GMBMFlag is "0", the received bitstream is transmitted to the second unpacking processor 913.

The second unpacking processor 913 unpacks the received bitstream in units of macroblocks and outputs the macroblock. The output unpacked bitstream is transmitted to the second dequantizer 922 of the dequantizer 920.

In addition, the flag detector 911 detects a flag SMFlag allocated in a slice unit, a flag SMFlag value is "0", and a flag GMBMFlag value allocated in a macroblock unit is "1". The received bitstream is transmitted to the third unpacking processor 914.

The third unpacking processor 914 unpacks the received bitstream in group units of the macro block and outputs the unpacked unit. Although unpacking of macroblocks is performed in macroblock units, macros included in macroblock groups because motion vectors are assigned only to the first macroblock among macroblocks included in macroblock groups. When unpacking the first macroblock among the blocks, unpacking of the allocated motion vectors is performed, but unpacking of the motion vectors is not performed when unpacking the remaining macroblocks included in the group of macroblocks. The output unpacked bitstream is transmitted to the third dequantizer 923 of the dequantizer 920.

The dequantization unit 920 dequantizes the unpacked bitstream in units of slices or macroblocks. In this case, the inverse quantization unit 920 is configured to perform separate inverse quantization processing for macroblocks included in a group of macroblocks and macroblocks not included in a group of macroblocks when inverse quantization of a macroblock unit occurs. do. Accordingly, the inverse quantizer 920 includes a first inverse quantizer 921, a second inverse quantizer 922, and a third inverse quantizer 923.

The first dequantizer 921 dequantizes the received unpacked bitstream in slice units. Accordingly, a motion vector dequantized in slice units is output. The bitstream dequantized in slice units is output to the motion compensator 930.

The second dequantizer 922 dequantizes the received unpacked bitstream in units of macro blocks. Accordingly, a motion vector dequantized in macroblock units is output. The bitstream dequantized in macroblock units is output to the motion compensation unit 930.

The third dequantizer 923 dequantizes the received unpacked bitstream in groups of macro blocks. Accordingly, a motion vector dequantized in group units of the macroblock is output. The bitstream dequantized in group units of the macroblock is output to the motion compensator 930.

When the dequantized data is received from the first dequantizer 921, the motion compensator 930 compensates the motion of all macro blocks included in the slice by using the dequantized motion vector in the slice unit. When the dequantized data is received from the second dequantizer 922, the motion compensator 930 compensates for the motion of the macro block by using the dequantized motion vector in macroblock units. When the inverse quantized data is received from the third inverse quantizer 923, the motion compensator 930 uses the inverse quantized motion vector in groups of macroblocks to move the macroblocks included in the macroblock group. To compensate.

FIG. 11 is a functional block diagram of an image encoding apparatus corresponding to the image decoding apparatus illustrated in FIG. 9. That is, FIG. 11 is a view illustrating a video encoding apparatus for allocating a flag indicating a motion vector allocation unit in a slice unit and a macroblock unit, and allocating a motion vector in a slice unit, a macroblock unit, or a group unit of a macroblock according to the allocated flag. This is a functional block diagram.

Referring to FIG. 11, the apparatus for encoding an image includes a slice-based motion estimator 110, a block-based motion estimator 1102, a macroblock grouping processor 1103, a motion estimation type selection and flag value determiner 1104, A quantization unit 1105 and a bitstream packing unit 1106.

The slice-based motion estimator 1101 is configured and operates in the same manner as the slice-based motion estimator 501 of FIG. 5. The block-based motion estimator 1102 is configured and operates in the same manner as the block-based motion estimator 801 of FIG. 8.

The grouping processor 1103 of the macro block is configured and operates in the same manner as the grouping processor 802 of the macro block of FIG. 8.

The motion estimation type selection and flag value determination unit 1104 compares the RD cost output from the slice-based motion estimation unit 1101 with the RD cost output from the grouping processing unit 1103 of the macroblock and outputs a small value. Select the type as the best type.

That is, if the RD cost output from the slice-based motion estimation unit 1101 is smaller than the RD cost output from the grouping processing unit 1103 of the macro block, the motion estimation type selection and flag value determination unit 1104 may slice-based motion estimation. The motion estimation type in the government unit 1101 is selected, and the value of the slice unit flag SMFlag is determined as "1". Next, the motion estimation type selection and flag value determination unit 1104 transmits the motion estimation result output from the slice-based motion estimation unit 1101 to the quantization unit 1105, and determines the slice unit flag value (SMFlag = 1). To the bitstream packing unit 1106.

On the other hand, if the RD cost output from the macroblock grouping processor 1103 is less than the RD cost output from the slice-based motion estimation unit 1101, the motion estimation type selection and flag value determination unit 1104 may determine the macroblock of the macroblock. The motion estimation type in the grouping processor 1103 is selected, and a flag SMFlag value of a slice unit is determined as "0". Next, the motion estimation type selection and flag value determination unit 1104 determines a flag GMBMFlag indicating a motion vector allocation unit allocated in macroblock units, based on the grouping processing result of the macroblock grouping processing unit 1103. That is, if the macroblock corresponding to the macroblock unit is not included in the group of the macroblock, the value of the flag GMBMFlag is determined as "0", and if the corresponding macroblock is included in the group of the macroblock, the flag GMBMFlag ) Is determined as "1".

The motion estimation type selection and flag value determination unit 1104 provides flag values indicating the determined motion vector allocation unit to the bitstream packing unit 1106 in slice units and macro block units, and slice-based motion according to the selected motion estimation type. The motion estimation result transmitted from the estimator 1101 and the block-based motion estimation result transmitted from the macroblock grouping processor 1103 are transmitted to the quantization unit 1105.

The quantization unit 1105 quantizes and outputs motion estimation image data transmitted from the motion estimation type selection and flag value determination unit 1104.

The bitstream packing unit 1106 outputs from the quantization unit 1105 based on the slice unit flag (SMFlag) and the macro block unit flag (GMBMFlag) value provided from the motion estimation type selection and flag value determination unit 1104. The bitstream of the quantized video signal is packed and output as shown in FIG. 10 (a) or 10 (b).

That is, when the slice unit flag SMFlag is "1", the bitstream packing unit 1106 packs the bitstream of the quantized video signal as shown in FIG. Accordingly, the bitstream packing unit 1106 outputs a bitstream in which one motion vector is allocated to one slice.

If there is a macro block in which the value of the slice unit flag SMFlag is "0" and the value of the macro block unit flag GMBMFlag is "1", the bitstream packing unit 1106 is shown in Fig. 10B. As described above, the bitstream of the quantized video signal is packed. Accordingly, the bitstream packing unit 1106 assigns independent motion vectors to macroblocks not included in the group of macroblocks, and the macroblocks included in the group of macroblocks store motion vectors representing the group of macroblocks. The bitstream is allocated to the first macroblock of the group of blocks and the remaining macroblocks included in the group of the macroblock are not assigned a motion vector.

12 is a flowchart illustrating an image decoding method according to another embodiment of the present invention.

When the bitstream is received, the image decoding method detects a flag indicating a motion vector allocation unit from the received bitstream (1201 and 1202). The flag is detected as in the flag detector 311 of FIG. 3 or the flag detector 911 of FIG. 9. In the image decoding method, the received bitstream is unpacked in a slice unit or a macroblock unit as described with reference to FIGS. 3 and 9 and dequantized (1203).

Next, the image decoding method compensates for the motion of the dequantized bitstream using the motion vector included in the dequantized bitstream, like the motion compensator 330 of FIG. 3 or the motion compensator 930 of FIG. 9. (1204).

FIG. 13 illustrates an example in which the image decoding method illustrated in FIG. 12 is applied when a flag indicating a motion vector allocation unit is detected in a slice unit.

When the bitstream is received, the image decoding method detects a flag indicating a motion vector allocation unit from the received bitstream (1301 and 1302). The flag is detected as in the flag detector 311 of FIG.

If the detected flag is a flag SMFlag in a slice unit and the value of the flag SMFlag is "0", the image decoding method performs unpacking and dequantization on the bitstream in units of macroblocks (1303 and 1304). ). In operation 1305, the image decoding method compensates for the motion of each macro block by using the motion vector dequantized in units of macro blocks.

On the other hand, if the detected flag is a flag SMFlag in slice units and the value of the flag SMFlag is "1", the image decoding method unpacks and dequantizes the bitstream in slice units (1306). In operation 1307, the image decoding method compensates for the motion of all the macroblocks included in the slice by using the motion vector dequantized in units of slices.

FIG. 14 illustrates an example in which the image decoding method illustrated in FIG. 12 is applied when a flag indicating a motion vector allocation unit is detected in macroblock units.

When the bitstream is received, the image decoding method detects a flag indicating a motion vector allocation unit from the received bitstream (1401 and 1402). The flag is detected as in the flag detector 311 of FIG.

If the detected flag is a macroblock unit flag (GMBMFlag) and the macroblock unit flag (GMBMFlag) has a value of "0", the video decoding method performs unpacking and dequantization on the bitstream in units of macroblocks. (1403, 1404). In operation 1405, the image decoding method compensates for the motion of each macro block by using the motion vector dequantized in units of macro blocks.

On the other hand, if the detected flag is a macroblock unit flag (GMBMFlag) and the macroblock unit flag (GMBMFlag) has a value of "1", the video decoding method unpacks and dequantizes the bitstream in units of macroblocks ( 1406). In this case, the video decoding method unpacks and dequantizes a motion vector in groups of macro blocks. This is because when a macroblock group is formed for a macroblock included in a corresponding slice, a motion vector representing a group of macroblocks is assigned to only the first macroblock among the macroblocks included in the macroblock group. Accordingly, bitstream unpacking and inverse quantization of the macroblock unit in step 1406 may be defined as bitstream unpacking and inverse quantization of the group unit of the macroblock.

In operation 1407, the image decoding method compensates for the motion of the macroblock included in the group of the macroblock by using the motion vector dequantized by the group of the macroblock.

FIG. 15 illustrates an example in which the image decoding method illustrated in FIG. 12 is applied when a flag indicating a motion vector allocation unit is detected in a slice unit and a macroblock unit.

When the bitstream is received, the image decoding method detects a flag indicating a motion vector allocation unit from the received bitstream (1501 and 1502). The flag is detected as in the flag detector 911 of FIG.

If the detected flag is a slice unit flag SMFlag and the slice unit flag SMFlag value is "1", the video decoding method unpacks and dequantizes the bitstream in slice units (1503 and 1504). Next, the image decoding method compensates for the motion of all the macroblocks included in the slice by using the dequantized motion vector in the slice unit (1505).

If the detected flag is a slice unit flag (SMFlag) and the slice unit flag (SMFlag) is "0", the video decoding method detects a flag indicating a motion vector allocation unit again in the received bitstream, If the flag GMBMFlag is detected and the value of the detected macroblock unit flag GMBMFlag is "0", the video decoding method unpacks and dequantizes the bitstream in units of macroblocks (1506 and 1507). In operation 1508, the image decoding method compensates for the motion of each macro block by using the motion vector dequantized in units of macro blocks.

If the detected flag is the flag SMFlag in the slice unit and the value of the flag SMFlag is "0", the video decoding method detects a flag indicating the motion vector allocation unit again in the received bitstream, and flags the macroblock unit. If (GMBMFlag) is detected and the value of the detected macroblock unit flag (GMBMFlag) is "1", the video decoding method unpacks and dequantizes the bitstream in units of macroblocks (1506 and 1509). However, at this time, the macroblock unit may be defined as a group unit of the macroblock. This is because one motion vector is allocated per group unit of a macroblock.

Next, the image decoding method compensates for the motion of each macro block included in the group of macro blocks by using the motion vector dequantized in group units of the macro block (1510).

An image encoding method corresponding to the image decoding method according to the present invention may be implemented based on FIGS. 5, 8, and 11. That is, a video encoding method may be implemented such that a flag indicating a motion vector allocation unit and a motion vector are allocated to a video signal input based on FIG. 5 in a slice unit as shown in FIGS. 4 (a) and 4 (b). Can be. Alternatively, an image encoding method may be implemented such that a flag indicating a motion vector allocation unit and a motion vector are allocated to a video signal input based on FIG. 8 in macroblock units as shown in FIG. 7. Alternatively, a flag indicating a motion vector allocation unit is allocated to a video signal input based on FIG. 11 in slice units and macroblock units as shown in FIGS. 10 (a) and 10 (b), and is divided into slice units or macros. An image encoding method may be implemented such that a motion vector is allocated in a block unit or a group unit of a macroblock.

The program for performing the image encoding and decoding method according to the present invention can be embodied as computer readable codes on a computer readable recording medium. Computer-readable recording media include all kinds of storage devices that store data that can be read by a computer system. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

As described above, the present invention reduces the number of bits of a motion vector included in a bitstream by allocating a motion vector in a slice unit, a macroblock unit, or a group unit of a macroblock during image encoding, and at the time of image decoding. By unpacking the motion vector included in the bitstream based on a flag indicating a motion vector allocation unit in units of slices, macroblocks, or groups of macroblocks, de-quantization, and using them for motion compensation, The compression rate of the image data can be improved. In particular, the present invention can further improve the compression rate of the image data when the still image lasts for a predetermined time.

Claims (8)

In the video decoding method, If a bitstream is received, detecting a flag indicating a motion vector allocation unit from the received bitstream; Unpacking and inverse quantizing the received bitstream in units of slices or macroblocks based on the detected flag; And Compensating for the motion of the dequantized bitstream using a motion vector included in the dequantized bitstream. The method of claim 1, wherein if a flag indicating the motion vector allocation unit is detected in a slice unit, The motion compensation step, Compensating for the motion of all macroblocks included in the slice using the dequantized motion vector in the slice unit or compensating for the motion of each macroblock using the motion vector dequantized in the macroblock unit. Image Decoding Method. The method of claim 1, wherein when a flag indicating the motion vector allocation unit is detected in macroblock units, The motion compensation step, Compensate for the motion of the macroblocks included in the group of macroblocks by using the dequantized motion vectors in groups of macroblocks, or use the motion vectors inversely quantized in each macroblock unit to perform motion for each macroblock. Image decoding method characterized in that the compensation. The method of claim 1, wherein if a flag indicating the motion vector allocation unit is detected in a slice unit and a macroblock unit, The motion compensation step, If the flag detected in the slice unit indicates that a motion vector is allocated in the slice unit, Compensating for the motions of all the macroblocks included in the slice by using the dequantized motion vector in the slice unit; If the flag detected in units of slices indicates that a motion vector is allocated in units of macro blocks, and the flag detected in units of macro blocks indicates that a motion vector is allocated in units of groups of macro blocks. Compensating for the motion of the macroblock included in the group of the macroblock by using the dequantized motion vector in the group of the macroblock; If the flag detected in units of slices indicates that a motion vector is allocated in units of macro blocks, and the flag detected in units of macro blocks indicates that a motion vector is allocated in units of macro blocks, And decoding the motion of each macro block using the motion vector dequantized in units of the macro blocks. In the video decoding apparatus, A bitstream unpacking unit configured to unpack the received bitstream in a slice unit or a macroblock unit based on a flag indicating a motion vector allocation unit included in the received bitstream when the bitstream is received; An inverse quantizer for inversely quantizing the unpacked bitstream in units of slices or units of macroblocks; A motion compensation unit for compensating for the motion of the dequantized bitstream in the dequantization unit, The flag indicating the motion vector allocation unit is allocated in a slice unit and / or a macro block unit, The motion compensation unit, Compensate for the motion of all macroblocks included in a slice using a dequantized motion vector in slices And a motion compensation for a macro block using a motion vector dequantized in macro block units. The method of claim 5, wherein the bitstream unpacking unit further includes a function of unpacking the received bitstream in units of macroblocks based on the flag. If the bitstream unpacking unit unpacks the received bitstream in groups of macro blocks, the dequantization unit further includes a function of inverse quantization of the unpacked bitstreams in groups of macro blocks. The motion compensator compensates for the motion of the macroblock included in the group of the macroblock using the motion vector dequantized in the group unit of the macroblock when the motion vector inversely quantized in the macroblock group is received. The video decoding apparatus further comprises. In the video encoding apparatus, A slice-based motion estimator for estimating a motion in a slice unit with respect to an input image signal and outputting an RD (rate-distortion) cost according to the motion estimation result in the slice unit and the motion estimation in the slice unit; A block-based motion estimator for estimating motion on a macroblock basis with respect to the input video signal and outputting an Aldi cost according to the motion estimation result of the macroblock unit and the motion estimation of the macroblock unit; Based on a result of comparing the Aldi cost output from the slice-based motion estimator and the Aldi cost output from the block-based motion estimator, a motion estimation type is selected to output motion estimation image data, and the motion estimation type is output to the selected motion estimation type. A motion estimation type selection and flag value determining unit configured to determine a flag value accordingly; A quantizer for quantizing the motion estimated image data output from the motion estimation type selection and flag value determining unit; And And a bitstream packing unit for packing a bitstream of the quantized image data output from the quantization unit based on the flag value provided from the motion estimation type selection and flag value determination unit. In the video encoding method, Estimating a motion in a slice unit with respect to an input image signal, and outputting an RD (rate-distortion) cost according to the motion estimation result in the slice unit and the motion estimation in the slice unit; Estimating motion on a macroblock basis with respect to the input video signal, and outputting an Aldi cost according to the motion estimation result of the macroblock unit and the motion estimation of the macroblock unit; Selecting a motion estimation type based on a result of comparing the Aldi cost according to the motion estimation in the slice unit with the Aldi cost according to the motion estimation in the macroblock unit, and determining a flag value according to the selected motion estimation type; Quantizing the motion estimated image data according to the selected motion estimation type; And packing a bitstream of the quantized image data based on the flag value.

Images