Method of encoding mode determination, method of motion estimation and encoding apparatus

• US 20050135484A1
• Filed: 06/07/2004
• Published: 06/23/2005
• Est. Priority Date: 12/18/2003
• Status: Abandoned Application

• Current Assignee: Daeyang Foundation, Samsung Electronics Co. Ltd.
• Original Assignee: Daeyang Foundation, Samsung Electronics Co. Ltd.
• Inventors: Han, Ki-hoon, Lee, Yung-lyul
• Application Number:
  US10/861,486
• Publication Number:
  US 20050135484A1
• Time in Patent Office: Days
• US Class Current: 375/240.160
• CPC Class Codes: H04N 19/109 among a plurality of temporal predictive coding modes
• CPC Class Codes: H04N 19/11 among a plurality of spatial predictive coding modes
• CPC Class Codes: H04N 19/147 according to rate distortion criteria rate-distortion as a criterion for motion estimation H04N19/567
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/186 the unit being a colour or a chrominance component
• CPC Class Codes: H04N 19/557 Motion estimation characterised by stopping computation or iteration based on certain criteria, e.g. error magnitude being too large or early exit
• CPC Class Codes: H04N 19/593 involving spatial prediction techniques
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

Motion estimation of a macro block in inter16×16, inter16×8, and inter8×16 modes is performed and a determination of whether to further perform motion estimation in a P8×8 mode is made. Motion estimation in P8×8 mode is either omitted or performed and one mode is determined according to a rate distortion cost of the respective modes. Spatial prediction encoding may then be performed or omitted based on comparing the rate distortion cost of the one mode with a predetermined value. Accordingly, by selectively omitting variable block motion estimation and spatial prediction encoding which are the most complicated operations in an H.264 encoder, determining an encoding mode is rapidly performed such that encoding speed increases.

Claims

1. A method of determining an encoding mode, comprising:

performing motion estimation of a macro block in an inter16×

16 mode, an inter16×

8 mode, and an inter8×

16 mode;

determining whether to further perform motion estimation in a P8×

8 mode;

according to the determination result, omitting or performing motion estimation in the P8×

8 mode and then selecting one mode from among the modes for which motion estimation has been performed; and

omitting or further performing spatial prediction encoding according to a rate-distortion cost of the selected mode and determining a final encoding mode.

2. The method of claim 1, wherein the determining of whether to further perform motion estimation in P8×

8 mode comprises;

for each of the Inter16×

16 mode, the Inter16×

8 mode, and the Inter8×

16 mode,

calculating a sum of absolute difference (SAD) value, which is a difference of a pixel value of a current picture and a pixel value of a previous picture, and

calculating MVcost, which is a difference value of a motion vector of the current picture and a motion vector of the previous picture; and

comparing a sum of the SAD and MVcost values (SAD+MVcost) of the respective modes and determining whether to further perform motion estimation in the P8×

8 mode based on the comparison.

3. The method of claim 2, wherein in the determining of whether to further perform motion estimation in the P8×

8 mode, if the (SAD+MVcost) value in the Inter16×

16 mode is smaller than the (SAD+MVcost) values in other modes, motion estimation in the P8×

8 mode is omitted.

4. The method of claim 2, wherein the omitting or performing motion estimation in the P8×

8 mode and then determining the one mode comprises;

if the (SAD+MVcost) value in the Inter16×

16 mode is smaller than the (SAD+MVcost) values in the other modes, calculating rate-distortion costs in the Inter16×

16 mode, the Inter16×

8 mode, the Inter8×

16 mode, and a SKIP mode; and

selecting one mode which has a smallest rate-distortion cost among the modes.

5. The method of claim 2, wherein the omitting or performing motion estimation in the P8×

8 mode and then determining the one mode comprises;

if the (SAD+MVcost) value in the Inter16×

16 mode is not smaller than the (SAD+MVcost) values in the other modes,

calculating rate-distortion costs in an Inter8×

8 mode, an Inter8×

4 mode, an Inter4×

8 mode and an Inter4×

4 mode,

selecting one mode which has a smallest rate-distortion cost from among the Inter8×

8 mode, the Inter8×

4 mode, the Inter4×

8 mode and the Inter4×

4 mode,

determining whether calculation of a rate-distortion cost in an IBLOCK mode is needed, and

calculating a rate-distortion cost in IBLOCK mode and selecting one mode from among the Inter8×

8 mode, the Inter 8×

4 mode, the Inter4×

8 mode, the inter 4×

4 mode, and the IBLOCK mode, if the calculation in the IBLOCK mode is needed.

6. The method of claim 5, wherein in the determining of whether calculation of the rate-distortion cost in the IBLOCK mode is needed,

if the rate-distortion cost in the selected mode is not less than one-fourth of the average value of the rate-distortion costs of all macro blocks previously encoded in spatial prediction mode, motion estimation is further performed in IBLOCK mode and the rate-distortion cost in the IBLOCK mode is calculated.

7. The method of claim 2, wherein the omitting or further performing spatial prediction encoding according to the rate-distortion cost value in the selected mode, and the determining of the final encoding mode comprises:

if the rate-distortion cost value in the selected mode is less than an average value of rate-distortion costs of all macro blocks previously encoded in the spatial prediction mode, motion estimation in Intra mode is not performed and the selected mode is determined as a final encoding mode; and

if the rate-distortion cost value in the determined mode is not less than the average value of rate-distortion costs of all macro blocks previously encoded in spatial prediction mode, motion estimation in Intra mode is further performed and the rate-distortion cost in the Intra mode is calculated, and by comparing the rate-distortion cost in the intra mode with the rate-distortion cost in the selected mode, a mode having a smaller value is determined as the final encoding mode.

8. The method of claim 7, wherein in the determining of a final encoding mode after performing motion estimation in the Intra mode is further performed, if the Intra mode is determined as the final encoding mode, the average value of rate-distortion costs of all macro blocks previously encoded in spatial prediction mode is updated based on the calculated rate-distortion cost in the intra mode.

9. The method of claim 8, wherein the updating of the average value is performed by multiplying the average value by the number of Intra macro blocks previously occurring, adding the multiplication result to the calculated rate-distortion cost in the Intra mode, and dividing the addition result by the number of Intra macro blocks previously occurring plus one.

10. A method of motion estimation of a macro block comprising:

performing motion estimation by using an entire macro block;

dividing the macro block into two in the horizontal direction or the vertical direction and then performing motion estimation by using each divided block;

dividing the macro block into 4 blocks of an identical size and performing motion estimation by using each divided block;

dividing each of the four divided blocks into two in the horizontal direction or the vertical direction and performing motion estimation by using each further divided block; and

dividing each of the four divided blocks into four and performing motion estimation by using each further divided block.

11. The method of claim 10, wherein in the performing of motion estimation by using the entire macro block, motion estimation is performed by using the entire macro block of a 16×

16 size.

12. The method of claim 10, wherein in the dividing of the macro block into two in the horizontal direction or the vertical direction and then performing motion estimation by using each divided macro block, motion estimation for the macro block is performed by using two 16×

8 blocks and then by using two 8×

16 blocks.

13. The method of claim 10, wherein in the dividing of the macro block into 4 blocks of an identical size and performing motion estimation by using each divided block, motion estimation for the macro block is performed by using two 16×

8 blocks and then by using two 8×

16 blocks.

14. An encoding apparatus comprising:

a DCT+Q performing unit which receives picture data and performs discrete cosine transform (DCT) and quantization;

a rate-distortion optimization unit which calculates a rate-distortion cost of the picture data and determines an encoding block mode to be used in encoding the picture, and transfers the determined block mode to the DCT+Q performing unit; and

a motion estimator and a motion compensator which by using a reference frame and the input picture, performs motion estimation and compensation and transfers the result to the DCT+Q performing unit.

15. The encoding apparatus of claim 14, wherein the rate-distortion optimization unit calculates a sum of absolute difference (SAD) value that is a difference of a pixel value of a current picture and a pixel value of a previous picture, in the input picture data, and MVcost that is a difference value of a motion vector of the current picture and a motion vector of the previous picture, and according to a value (SAD+MVcost), omits motion estimation in spatial prediction mode.

16. A method of determining an encoding mode, comprising:

performing motion estimation of a macro block in a plurality of modes including an inter16×

16 mode, an inter16×

8 mode, and an inter8×

16 mode;

determining a SAD value, an MVcost value and a sum of the SAD value and the MVcost value (SAD+MVcost) for each of the plurality of modes, the SAD value being a difference of a pixel value of a current picture and a pixel value of a previous picture and the MVcost value being a difference value of a motion vector of the current picture and a motion vector of a previous picture;

calculating a rate distortion cost (RDcost) in each of the plurality of modes and in a skip mode, if the (SAD+MVcost) for the inter 16×

16mode is smaller than the (SAD+MVcost) corresponding to the other ones of the plurality of modes;

selecting the mode having the smallest RDcost from among the plurality of modes and the skip mode; and

if the RDcost of the selected mode is less than M, determining the selected mode as the encoding mode, where M is a mean RDcost value of a plurality Intra macro blocks encoded in a spatial prediction mode and occurring in previous frames.

17. The method of claim 16, wherein, if the RDcost of the selected mode is not less than M:

calculating an RDcost in Intra mode; and

determining the one of the Intra mode and the selected mode having the smallest RDcost as the encoding mode.

18. A method of determining an encoding mode, comprising:

performing motion estimation of a macro block in a first plurality of modes comprising an inter16×

16 mode, an inter16×

8 mode, and an inter8×

16 mode;

determining a SAD value, an MVcost value and a sum of the SAD value and the MVcost value (SAD+MVcost) for each of the first plurality of modes, the SAD value being a difference of a pixel value of a current picture and a pixel value of a previous picture and the MVcost value being a difference value of a motion vector of the current picture and a motion vector of a previous picture;

performing motion estimation in a each of a second plurality of modes, if the (SAD+MVcost) for the inter16×

16 mode is not smaller than the (SAD+MVcost) corresponding to the other ones of the first plurality of modes, the second plurality of modes comprising an Inter8×

8 mode, an Inter8×

4mode, an Inter4×

8mode and an inter 4×

4mode;

calculating an RDcost value for each of the second plurality of modes;

selecting the mode from among the second plurality of modes having the smallest RDcost, if the RDcost of the mode having the smallest RDcost is less than M/4, where M is a mean RDcost value of a plurality of Intra macro blocks occurring in previous frames and encoded in a spatial prediction mode; and

if the RDcost of the selected mode is less than M, determining the selected mode as the encoding mode,

19. The method of claim 18, wherein, if the RDcost of the selected mode is not less than M:

calculating an RDcost in Intra mode; and

determining the one of the Intra mode and the selected mode having the smallest RDcost as the encoding mode.