Distributed video decoder and distributed video decoding method

• US 20100177893A1
• Filed: 10/14/2009
• Published: 07/15/2010
• Est. Priority Date: 01/09/2009
• Status: Active Grant

• Current Assignee: Sungkyunkwan University Foundation For Corporate Collaboration (SungKyunKwan University)
• Original Assignee: Sungkyunkwan University Foundation For Corporate Collaboration (SungKyunKwan University)
• Inventors: Do, Tae-Won, Jeon, Byeung-Woo, Shim, Hiuk-Jae
• Granted Patent: US 8,599,929 B2
• Time in Patent Office: Days
• US Class Current: 380/239
• CPC Class Codes: H04N 19/395 involving distributed video coding [DVC], e.g. Wyner-Ziv video coding or Slepian-Wolf video coding
• CPC Class Codes: H04N 19/44 Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
• CPC Class Codes: H04N 19/51 Motion estimation or motion compensation

Abstract

A distributed video decoder and a distributed video decoding method are provided. The distributed video decoder includes a key picture decoding unit which reconstructs a key picture received from an encoder, a hash information selecting unit which determines necessary hash information using the reconstructed key picture and/or generated previous side information, and a side information generating unit which generates side information using the reconstructed key picture and/or hash information selected based on information determined by the hash information selecting unit. The distributed video decoder further includes a channel code decoding unit which estimates a quantized value by using a parity bit received from the encoder and the generated side information, and a video reconstructing unit which reconstructs a current picture as a decoding target by using the quantized value estimated by the channel code decoding unit and the generated side information.

Claims

1. A distributed video decoder, comprising:

a key picture decoding unit which reconstructs a key picture received from an encoder;

a hash information selecting unit which determines necessary hash information using the reconstructed key picture and/or generated previous side information; and

a side information generating unit which generates side information using the reconstructed key picture and/or hash information selected based on information determined by the hash information selecting unit.

2. The distributed video decoder of claim 1, further comprising:

a channel code decoding unit which estimates a quantized value by using a parity bit received from the encoder and the generated side information; and

a video reconstructing unit which reconstructs a current picture as a decoding target by using the quantized value estimated by the channel code decoding unit and the generated side information.

3. The distributed video decoder of claim 1, wherein the hash information selecting unit computes estimation reliability of current side information to be reconstructed by using the reconstructed key picture and/or the generated previous side information and transmits hash selection information for selecting necessary hash information by predetermined hash selection criterion information.

4. The distributed video decoder of claim 3, wherein the hash information selecting unit comprises:

a motion estimation reliability measuring unit which measures reliability of a motion vector within the current side information to be reconstructed based on the reconstructed key picture and/or the generated previous side information; and

a hash information determining unit which determines hash information necessary for generating side information with reduced noise based on the motion estimation reliability measured by the motion estimation reliability and the predetermined hash selection criterion information.

5. The distributed video decoder of claim 4, wherein the motion estimation reliability measuring unit comprises:

at least one of a temporal similarity measuring unit and a spatial similarity measuring unit, the temporal similarity measuring unit measuring temporal similarity between the reconstructed key pictures or between the reconstructed key picture and the generated previous side information, the spatial similarity measuring unit measuring spatial similarity between a certain pixel within the reconstructed key picture and/or the generated side information and a neighboring pixel; and

a final reliability determining unit which finally determines motion estimation reliability based on the temporal similarity and/or the spatial similarity.

6. The distributed video decoder of claim 5, wherein the temporal similarity measuring unit measures the temporal similarity by using a sum of pixel value differences between pixels within corresponding coding units of the reconstructed key pictures.

7. The distributed video decoder of claim 5, wherein the temporal similarity measuring unit measures the temporal similarity by comparing a motion vector between the reconstructed key pictures and a motion vector between the reconstructed key picture and the generated previous side information.

8. The distributed video decoder of claim 5, wherein the spatial similarity measuring unit measures the spatial similarity by using a pixel value difference between the certain pixel and the neighboring pixel.

9. The distributed video decoder of claim 4, wherein the hash information determining unit comprises:

a hash information position determining unit which selects a portion which needs hash information by using the motion estimation reliability and the predetermined hash selection criterion information;

a hash information type determining unit which determines a hash information type used in the selected portion; and

a hash information amount determining unit which determines a hash information amount of the determined type corresponding to the selected portion.

10. The distributed video decoder of claim 9, wherein the hash information position determining unit selects a picture which needs hash information at the time of side information generation or a position of a portion which needs hash information corresponding to a coding unit within a picture, the coding unit including a macroblock or a block of a predetermined size,

the determined hash information type includes at least one of a pixel value, a sum of pixel values within a block, an average of pixel values within a block, a standard deviation of pixel values within a block, boundary information, DC information, and AC information, and

the hash information amount determining unit outputs an overall hash information according to the selected position and the determined type or outputs hash information of a smaller amount.

11. The distributed video decoder of claim 2, wherein the side information generating unit comprises:

a has information determining unit which determines whether or not hash information is received from the encoder;

a first side generating unit which generates side information using the reconstructed key picture and the hash information;

a second side generating unit which generates side information using the reconstructed key picture; and

a final side information generating unit which stores final side information to be transmitted to the channel code decoding unit and the video decoding unit.

12. The distributed video decoder of claim 11, wherein the hash information determining unit determines whether or not hash information is received from the hash information generating unit, transmits the reconstructed key picture and the hash information to the first side information generating unit when the hash information is received, and transmits the reconstructed key picture to the second side information generating unit when the hash information is not received.

13. The distributed video decoder of claim 2, further comprising:

a transform unit which moves the reconstructed key pictures to a transform area when a decoding area of pictures to be reconstructed is a transform area other than a pixel area; and

an inverse transform unit which moves reconstructed pictures output from the video reconstructing unit to the pixel area.

14. A distributed video decoder, comprising:

a frame buffer which stores reconstructed key pictures reconstructed from key pictures received from an encoder; and

a decoding unit which selectively requests the encoder to transmit hash information with respect to side information which is determined as a noise is difficult to be removed by using a parity received from the encoder.

15. The distributed video decoder of claim 14, wherein the decoding unit selectively requests the encoder to transmit hash information with respect to side information which is determined as a noise is difficult to be removed by using a parity received from the encoder, generates side information using the reconstructed key picture and/or hash information selected by the encoder, and reconstructs a current picture as a decoding target by using a parity bit received from the encoder and the side information.

16. A distributed video encoder, comprising:

a key picture decoding unit which reconstructs a key picture received from an encoder;

a side information generating unit which generates side information using the reconstructed key picture;

a channel code decoding unit which estimates a quantized value by using a parity bit received from the encoder and the generated side information;

an intra coding selecting unit which estimates an amount of a noise included in the generated side information and selects side information or a coding unit within side information which is estimated having a lot of noise; and

a video reconstructing unit which reconstructs a current picture as a decoding target by using the quantized value estimated by the channel code decoding unit and the side information.

17. A distributed video encoder, comprising:

a key picture decoding unit which reconstructs a key picture received from an encoder;

a side information generating unit which generates side information using the reconstructed key picture;

a parity request skip selecting unit which determines an image quality of the generated side information;

a channel code decoding unit which skips a parity bit request to the encoder and estimates a quantized value by using the side information when it is determined the image quality of the generated side information is better than a predetermined criterion; and

a video reconstructing unit which reconstructs a current picture as a decoding target by using the quantized value estimated by the channel code decoding unit and the side information.

18. A distributed video decoding method, comprising:

reconstructing a key picture received from an encoder;

determining necessary hash information using the reconstructed key picture and/or generated previous side information; and

generating side information using the reconstructed key picture and/or the determined hash information.

19. The distributed video method of claim 18, further comprising:

estimating a quantized value by using a parity bit received from the encoder and the generated side information; and

reconstructing a current picture as a decoding target by using the estimated quantized value and the generated side information.

20. The distributed video method of claim 19, wherein the determining of necessary hash information comprises:

computing estimation reliability of current side information to be reconstructed by using the reconstructed key picture and/or the generated previous side information; and

transmitting hash selection information for selecting necessary hash information by the estimation reliability and predetermined hash selection criterion information.

21. The distributed video method of claim 20, wherein the computing of the estimation reliability comprises:

at least one of measuring temporal similarity between the reconstructed key pictures or between the reconstructed key picture and the generated previous side information and measuring spatial similarity between a certain pixel within the reconstructed key picture and/or the generated side information and a neighboring pixel; and

finally determining motion estimation reliability based on at least one of the temporal similarity and the spatial similarity.

22. The distributed video method of claim 21, wherein the temporal similarity is measured by using a sum of pixel value differences between pixels within corresponding coding units of the reconstructed key pictures, a sum of pixel value differences between corresponding pixels within corresponding coding units of the reconstructed key picture and the generated previous side information, or a difference between a motion vector between the reconstructed key pictures and a motion vector between the reconstructed key picture and the generated previous side information.

23. The distributed video method of claim 21, wherein the spatial similarity is measured by using a pixel value difference between a certain pixel and a neighboring pixel within the reconstructed key picture and/or the generated previous side information.

24. The distributed video method of claim 20, wherein the determining of necessary hash information comprises:

determining a position of a portion which needs hash information corresponding to side information or a coding unit within side information which needs hash information and determining a has information position by using the motion estimation reliability and the predetermined hash selection criterion information, the coding unit including a macroblock or a block of a predetermined size;

determining a hash information type in the side information or the coding unit within the side information which needs the hash information; and

determining a hash information amount after the hash information position and type are determined.

25. The distributed video method of claim 24, wherein the hash information type includes at least one of a pixel value, a sum of pixel values within a block, an average of pixel values within a block, a standard deviation of pixel values within a block, boundary information, DC information, and AC information.

26. The distributed video method of claim 18, wherein the generating of the side information comprises:

determining whether or not the hash information is received;

generating first side information using the reconstructed key picture and the hash information when it is determined that the hash information is received;

generating second side information using the reconstructed key picture when it is determined that the hash information is not received; and

determining final side information based on the first side information and the second side information.