Back to results
US 5,845,013 AGranted
Region-based texture coding method and decoding method, and corresponding systems
AssigneeU.S. Philips Corporation
InventorsLionel Bouchard, Regine Askenatzis
Life of this patent
Estimating the term…
Abstract
The invention relates to an image sequence coding method in which images are segmented and coded with respect to their contours and textures. The texture coding step is carried out by means of a lattice vector quantization method adapted to region-based coding schemes where regions have arbitrary shapes, for applications allowing to reach very low bit rates while keeping a good image quality.
Claims8
A method of coding images of an image sequence, the method comprising the following steps:
- (A) segmenting each image in the image sequence into homogeneous regions comprised of textures and contours;
- (B) coding the contours of said regions;
- and
- (C) coding the textures of said regions;
- wherein, for each region, said texture coding step comprises:
- (a) building a smallest rectangle surrounding the region;
- (b) splitting the rectangle into 2.times.2 blocks;
- (c) storing transformed coefficients for each of the 2.times.2 blocks in vectors having sizes defined as 1, 2, 3 or 4;
- (d) selecting a codebook for each vector, the selecting step selecting a codebook for each vector that corresponds to a size of the vector;
- and
- (e) performing lattice vector quantization on each vector in accordance with the codebook selected in the selecting step.
8 more elements
A system for coding images of an image sequence, the system comprising:
- segmentation means for defining homogeneous regions in each image in the image sequence;
- contour coding means for coding contours of the regions of each image;
- and
- texture coding means for coding textures of the regions of each image, said texture coding means comprising a processor which, for each region, (a) builds a smallest rectangle surrounding the region, (b) splits said rectangle into 2.times.2 blocks, (c) stores transformed coefficients for each of the 2.times.2 blocks in vectors having sizes defined as 1, 2, 3 or 4, (d) selects a codebook for each vector corresponding to a size of the vector, and (e) performs lattice vector quantization on each vector in accordance with a selected codebook.
1 more element
Dependencies are inferred from the claim text.