VIDEO ENCODING METHOD AND DEVICE AND DECODING METHOD AND DEVICE

• US 20130188730A1
• Filed: 09/28/2011
• Published: 07/25/2013
• Est. Priority Date: 09/28/2010
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Hong, Yoon-Mi, Han, Woo-Jin, Chen, Jianle, Cheon, Min-Su
• Granted Patent: US 9,350,997 B2
• Time in Patent Office: Days
• US Class Current: 375/240.18
• CPC Class Codes: G06F 17/147 Discrete orthonormal transforms, e.g. discrete cosine transform, discrete sine transform, and variations therefrom, e.g. modified discrete cosine transform, integer transforms approximating the discrete cosine transform G06F17/145 takes precedence
• CPC Class Codes: H04N 19/126 Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/60 using transform coding
• CPC Class Codes: H04N 19/625 using discrete cosine transform [DCT]

Abstract

A video encoding method, a video encoding apparatus, a video decoding method, and a video decoding apparatus, the video encoding method including producing a fast transform matrix based on a transform matrix which is used for frequency transformation on a predetermined-size block; producing a transformed block by transforming the predetermined-size block by using the fast transform matrix; and performing scaling with respect to the transformed block to correct a difference between the transform matrix used for the frequency transformation and the fast transform matrix.

Claims

1. A video encoding method comprising:

producing a fast transform matrix based on a first transform matrix which is used for performing a frequency transformation on block which has a predetermined size;

producing a transformed block by transforming the block which has the predetermined size by using the fast transform matrix; and

performing scaling with respect to the transformed block in order to correct a difference between the first transform matrix and the fast transform matrix.

2. The video encoding method of claim 1, wherein the producing of the fast transform matrix comprises producing the fast transform matrix by substituting values based on a trigonometrical function from among elements of the first transform matrix with dyadic rationals.

3. The video encoding method of claim 1, wherein the producing the fast transform matrix comprises producing the fast transform matrix by rounding up, to a respective nearest integer, a value obtained by multiplying each of the elements of the first transform matrix by a predetermined constant.

4. The video encoding method of claim 1, the performing of the scaling comprises one of a) using a vector product operation of an N×

N transformed block and a predetermined scaling matrix; and

b) multiplying the N×

N transformed block by a predetermined scaling constant.

5. The video encoding method of claim 4, wherein each of the scaling matrix and the scaling constant is determined based on a quantization parameter.

6. The video encoding method of claim 5, wherein, when QP denotes the quantization parameter, six scaling matrices corresponding to respective cases where the quantization parameter QP is respectively labeled as zero, one, two, three, four, and five and a scaling constant are set, and one is selected from among the six scaling matrices and the scaling constant based on a value of (QP mod 6).

7. The video encoding method of claim 4, wherein six scaling matrices are set for an M×

M block based on the quantization parameter, and a scaling matrix obtained by multiplying each of the elements of the six scaling matrices set for the M×

M block by M/N is used as a scaling matrix for an N×

N block which has a size which is smaller than or greater than a size of the M×

M block, wherein M is an integer.

8. A video decoding method comprising:

receiving a quantized transformed block which has a predetermined size;

performing descaling with respect to the quantized transformed block in order to correct a difference between a first inverse-transform matrix which is used for performing a frequency inverse-transformation on the quantized transformed block and a fast inverse-transform matrix which is produced based on the first inverse-transform matrix; and

producing an inversely-transformed block by inversely transforming a descaled transformed block which is obtained by the descaling by using the fast inverse-transform matrix.

9. The video decoding method of claim 8, wherein the fast inverse-transform matrix is produced by substituting values based on a trigonometrical function from among elements of the first inverse-transform matrix which is used for performing a one-dimensional inverse discrete cosine transformation with dyadic rationals.

10. The video decoding method of claim 8, wherein the fast inverse-transform matrix is produced by rounding up, to a respective nearest integer, a value obtained by multiplying each of the elements of the first inverse-transform by 2ⁿ, wherein n is an integer.

11. The video decoding method of claim 8, wherein the performing of the descaling comprises one of a) using a vector product operation with respect to a quantized N×

N transformed block and a predetermined descaling matrix; and

b) multiplying the N×

N inversely-transformed block by a predetermined descaling constant.

12. The video decoding method of claim 11, wherein each of the descaling matrix and the descaling constant is determined based on a quantization parameter.

13. The video decoding method of claim 12, wherein, when QP denotes the quantization parameter, six descaling matrices corresponding to respective cases where the quantization parameter QP is respectively labeled as zero, one, two, three, four, and five and a descaling constant are set, and one is selected from among the six descaling matrices and the descaling constant based on a value of (QP mod 6).

14. The video decoding method of claim 13, wherein six descaling matrices are set for an M×

M block based on the quantization parameter, and a descaling matrix obtained by multiplying each of the elements of the six scaling matrices set for the M×

M block by M/N is used as a descaling matrix for an N×

N block which has a size which is smaller than or greater than a size of the M×

M block.

15. The video decoding method of claim 13, wherein, when i=(QP mod 6), the descaling constant DQMati comprises DQMat0=40, DQMat1=45, DQMat2=51, DQMat3=57, DQMat4=64, and DQMat5=72.

16. A video encoder apparatus, comprising:

a transformer which produces a fast transform matrix based on a first transform matrix which is used for performing a frequency transformation on a first block, and which transforms the first block by using the fast transform matrix in order to produce a fast-transformed first block; and

a quantizer which performs scaling with respect to the fast-transformed first block such that a difference between the fast-transformed first block and a result of transforming the first block by using the first transform matrix is reduced.

17. The video encoder apparatus of claim 16, further comprising:

a predictor which receives an input image, divides the received input image into at least one block, and generates a prediction block by performing one of inter prediction and intra prediction on the at least one block; and

a subtracter which subtracts the prediction block from corresponding data which is included in the input image in order to produce a residual, and provides the residual to the transformer.

18. The video encoder apparatus of claim 17, further comprising an entropy encoder which receives a result of the scaling which is performed with respect to the fast-transformed first block and which generates a bitstream by performing variable length encoding on the received result of the scaling.

19. The video encoder apparatus of claim 16, wherein the transformer produces the fast transform matrix by substituting values based on a trigonometrical function from among elements of the first transform matrix with dyadic rationals.

20. The video encoder apparatus of claim 16, wherein the transformer produces the fast transform matrix by rounding up, to a respective nearest integer, each value obtained by multiplying each of the elements of the first transform matrix by a predetermined constant.

21. The video encoder apparatus of claim 16, wherein the quantizer performs the scaling by using a vector product operation of an N×

N transformed block and a predetermined scaling matrix.

22. The video encoder apparatus of claim 16, wherein the quantizer performs the scaling by multiplying an N×

N transformed block by a predetermined scaling constant.

23. A video decoder apparatus, comprising:

receiving a quantized transformed block which has a predetermined size;

a dequantizer which performs descaling with respect to a received quantized transformed first block which has been generated by transforming a first block using a first transform matrix; and

an inverse transformer which inversely transforms a result of the descaling by using a fast inverse-transform matrix which is produced based on a first inverse-transform matrix which relates to the first transform matrix,

wherein the dequantizer performs the scaling such that a difference between a respective at least one element of the fast inverse-transform matrix and a corresponding at least one element of the first inverse-transform matrix is reduced.

24. The video decoder apparatus of claim 23, further comprising an entropy decoder which extracts prediction mode information, reference picture information, and residual information relating to a current block to be decoded and which provides a result of the extracting to the dequantizer.

25. The video decoder apparatus of claim 23, wherein the fast inverse-transform matrix is produced by substituting values based on a trigonometrical function from among elements of the first inverse-transform matrix with dyadic rationals.

26. The video decoder apparatus of claim 23, wherein the fast inverse-transform matrix is produced by rounding up, to a respective nearest integer, each value obtained by multiplying each of the elements of the first inverse-transform by 2ⁿ, wherein n is an integer.

27. The video decoder apparatus of claim 23, wherein the dequantizer performs descaling by using a vector product operation with respect to a quantized N×

N transformed block and a predetermined descaling matrix.

28. The video decoder apparatus of claim 23, wherein the dequantizer performs descaling by multiplying a quantized N×

N inversely-transformed block by a predetermined descaling constant.