Inter-Layer Motion Prediction Method

• US 20090225846A1
• Filed: 12/18/2006
• Published: 09/10/2009
• Est. Priority Date: 01/05/2006
• Status: Active Grant

• Current Assignee: Thomson Licensing (Vantiva SA)
• Original Assignee: Thomson Licensing (Vantiva SA)
• Inventors: Chen, Ying, Francois, Edouard, Lopez, Patrick, Bottreau, Vincent, Vieron, Jérome
• Granted Patent: US 8,446,956 B2
• Time in Patent Office: Days
• US Class Current: 375/240.160
• CPC Class Codes: H04N 19/16 for a given display mode, e.g. for interlaced or progressive display mode
• CPC Class Codes: H04N 19/187 the unit being a scalable video layer
• CPC Class Codes: H04N 19/31 in the temporal domain
• CPC Class Codes: H04N 19/33 in the spatial domain
• CPC Class Codes: H04N 19/52 by predictive encoding
• CPC Class Codes: H04N 19/57 Motion estimation characterised by a search window with variable size or shape

Abstract

The invention relates to a method for generating for at least one macroblock of a high resolution image at least one motion predictor from motion data associated to macroblock(s) of a low resolution image. The method comprises the following steps: —defining a virtual structure made up of non overlapping macroblocks, called virtual base layer macroblocks; —identifying for the high resolution macroblock corresponding virtual base layer macroblocks in the virtual image on the basis of the location of the high resolution macroblock within the high resolution image and on the basis of the ratios of low resolution and high resolution images dimensions; —deriving, for each identified virtual base layer macroblocks, motion data from motion data associated to at least one macroblock of a corresponding low resolution macroblocks pair on the basis of the field/frame mode of the corresponding low resolution macroblocks pair and on the basis of the field/frame mode of the high resolution macroblock; —resampling motion data associated to the corresponding virtual base layer macroblocks so as to generate the at least one motion predictor.

Claims

1. Method as part of a process of scalable video coding or decoding for generating for at least one macroblock of a high resolution image, called high resolution macroblock, at least one motion predictor from motion data associated to macroblock(s) of a low resolution image, called low resolution macroblock(s), comprising the following steps:

defining a virtual structure made up of non overlapping macroblocks, called virtual base layer macroblocks, said virtual structure having same dimensions as said low resolution image;

identifying for said high resolution macroblock corresponding virtual base layer macroblocks in said virtual image on the basis of the location of said high resolution macroblock within said high resolution image and on the basis of the ratios of low resolution and high resolution images dimensions;

associating to each identified virtual base layer macroblocks motion data, said motion data being deduced from motion data associated to at least one macroblock of a corresponding low resolution macroblocks pair on the basis of the field/frame mode of said corresponding low resolution macroblocks pair and on the basis of the field/frame mode of said high resolution macroblock;

resampling motion data associated to said corresponding virtual base layer macroblocks so as to generate said at least one motion predictor.

2. Method according to claim 1, wherein, if said high resolution macroblock is in frame mode and if said corresponding low resolution macroblocks pair is in field mode, the deriving step comprises the steps of:

upsampling vertically by a ratio of two said low resolution macroblocks pair made up of a top macroblock and a bottom macroblock, each upsampled macroblock being made up of an up part and a down part; and

merging motion data of the up part of the upsampled top macroblock with motion data of the up part of the upsampled bottom macroblock and merging motion data of the down part of the upsampled top macroblock with motion data of the down part of the upsampled bottom macroblock.

3. Method according to claim 1, wherein, if said high resolution macroblock is in field mode and if said corresponding low resolution macroblocks pair is in frame mode, the deriving step comprises a step for downsampling vertically by a ratio of two motion data associated to said corresponding low resolution macroblocks pair.

4. Device for coding a sequence of high resolution images made up of macroblocks, called high resolution macroblocks, and a sequence of low resolution images, called low resolution macroblocks, said device comprising:

first coding means for coding said low resolution images, said first coding means generating motion data for macroblocks of said low resolution images and a base layer data stream;

inheritance means for deriving motion data for at least one macroblock of a high resolution image, called high resolution macroblock from motion data of macroblocks a low resolution image, called low resolution macroblocks; and

second coding means for coding said high resolution images using said derived motion data, said second coding means generating an enhancement layer data stream;

wherein the inheritance means comprise;

means for defining a virtual structure made up of non overlapping macroblocks, called virtual base layer macroblocks, said virtual structure having same dimensions as said low resolution image;

means for identifying for said high resolution macroblock corresponding virtual base layer macroblocks in said virtual image on the basis of the location of said high resolution macroblock within said high resolution image and on the basis of the ratios of low resolution and high resolution images dimensions;

means for deriving, for each identified virtual base layer macroblocks, motion data from motion data associated to at least one macroblock of a corresponding low resolution macroblocks pair on the basis of the field/frame mode of said corresponding low resolution macroblocks pair and on the basis of the field/frame mode of said high resolution macroblock; and

means for resampling motion data associated to said corresponding virtual base layer macroblocks so as to generate said at least one motion predictor.

5. Device according to claim 4, wherein said first coding means are an MPEG-4 AVC video encoder.

6. Device for decoding at least a coded sequence of high resolution images, the coded images arising in the form of a data stream, each image being divided in non overlapping macroblocks, said device comprising:

first decoding means for decoding at least a first part of said data stream in order to generate low resolution images and motion data for macroblocks of said low resolution image;

inheritance means for deriving motion data for at least one macroblock of a high resolution image from motion data of macroblocks of a low resolution image; and

second decoding means for decoding at least a second part of said data stream using said derived motion data in order to generate high resolution images;

wherein the inheritance means comprise;

means for defining a virtual structure made up of non overlapping macroblocks, called virtual base layer macroblocks, said virtual structure having same dimensions as said low resolution image;

means for identifying for said high resolution macroblock corresponding virtual base layer macroblocks in said virtual image on the basis of the location of said high resolution macroblock within said high resolution image and on the basis of the ratios of low resolution and high resolution images dimensions;

means for deriving, for each identified virtual base layer macroblocks, motion data from motion data associated to at least one macroblock of a corresponding low resolution macroblocks pair on the basis of the field/frame mode of said corresponding low resolution macroblocks pair and on the basis of the field/frame mode of said high resolution macroblock; and

means for resampling motion data associated to said corresponding virtual base layer macroblocks so as to generate said at least one motion predictor.

7. Device according to claim 6, wherein said first decoding means are an MPEG-4 AVC video decoder.