Method for the encoding of data for assistance in the reconstruction of a sub-sampled moving electronic image

• US 4,980,764 A
• Filed: 06/21/1989
• Issued: 12/25/1990
• Est. Priority Date: 06/24/1988
• Status: Expired due to Fees

• Current Assignee: Etat Francais, Telediffusion De France SA
• Original Assignee: Etat Francais, Telediffusion De France SA
• Inventors: Henot, Jean-Pierre M.
• Primary Examiner(s): Kostak, Victor R.
• Application Number:
  US07/369,258
• Time in Patent Office: 552 Days
• Field of Search: 358/133, 358/135, 358/136, 358/138, 358/140, 358/141, 358/12, 358/13, 358/105
• US Class Current: 348/397.1
• CPC Class Codes: G06T 9/008 Vector quantisation
• CPC Class Codes: H04N 19/53 Multi-resolution motion estimation; Hierarchical motion estimation
• CPC Class Codes: H04N 19/96 Tree coding, e.g. quad-tree coding
• CPC Class Codes: H04N 7/0155 using pixel blocks

Abstract

A method for encoding data for assistance in the reconstruction of sub-sampled moving electronic images, particularly HDTV images transmitted by MAC channel, optimizing the representation of the assistance data while minimizing the corresponding throughput of data through compression of the image signal by segmentation of the image into homogeneous blocks of different sizes. An imbricated, hierarchical encoding is achieved on the image blocks, for assigning to the blocks a choice of processing, on the one hand, and a motion vector, on the other hand. When the sub-sampling/reconstruction processing chosen is of the motion compensation type (way 4) and wherein assistance data are generated under the control of a sequencer locked into a pre-determined image scanning procedure, segmentation information is selectively generated solely for the non-elementary blocks, and information on choice of motion vector is selectively generated only during the occurrence of a block with optimal processing by motion compensation.

Claims

1. A method for encoding assistance data for use in the reconstruction of an electronic image belonging to a sequence of images using a compressed image signal, wherein the compression of the image signal comprises the steps of:

dividing the electronic image into elementary image blocks containing a fixed number of pixels,

selecting an optimal processing operation from among several concurrent compression processing operations achieved in parallel for each elementary image block, the processing operations comprising at least one compression processing operation by image filtering and at least one compression processing operation by motion compensation,

segmenting the image into substantially homogeneous superblocks of variable size i×

i for which all elementary image blocks are defined collectively, and classifying each superblock according to the level of segmentation from a minimum size containing a single elementary image block, each level corresponding to the regrouping of four lower level adjacent blocks, the encoding method comprising the steps of;

generating for at least two levels of segmentation a first code for each superblock representing the homogeneity of the superblock, the compression processing employed, and the level of segmentation,

generating a second code for each superblock representing a motion vector where motion compensation compression processing is employed and the level of segmentation, and

multiplexing the codes together under the control of a sequencing means locked into a pre-determined image scanning procedure such that segmentation information is retained only for superblocks containing more than one elementary image block, and motion vector information is expressed only for blocks where motion compensation compression processing is employed.

2. The encoding method of claim 1 wherein the first code generating step comprises:

associating with each superblock at each level of segmentation a bit signifying the compression processing employed and another bit signifying the validity of such compression processing,

exploring iteratively the validity of all blocks within each superblock, and

retaining the code for the highest level superblock in which the bit signifying the compression processing employed is uniform for all blocks forming the superblock.

3. The encoding method of claim 1 wherein the second code generating step comprises:

associating with each elementary block a bit group signifying the value of the corresponding motion vector,

exploring iteratively the bit groups of all blocks within each superblock, and

retaining the code for the highest level superblock in which the bit group signifying the motion vectors in each block is uniform.

4. The encoding method of claim 3 wherein the second code generating step further comprises:

initializing a counter associated with a first elementary block having a motion vector within each superblock,

during said iterative exploration incrementing said counter at each occurrence of a same value motion vector within each superblock, and

examining the counter at the end of each iterative exploration to determine the homogeneity of the superblock.

5. The encoding method of claim 1 wherein the multiplexing step comprises:

forming a sentence of a variable number of words for each superblock at the maximum level of homogeneous segmentation, each sentence including at least one word selected from the group consisting of;

a first word identifying the segmentation level of the superblock,

a second word identifying the type of compression processing operation associated with the superblock,

a third word identifying the segmentation of motion within the superblock, and

a fourth word identifying the motion vector of the superblock.

6. The encoding method of claim 5 further comprising the step of transmitting the first or third word only for superblocks containing more than one elementary image block.

7. The encoding method of claim 5 further comprising the step of transmitting the fourth word only for blocks where motion compensation compression processing is employed.

8. The encoding method of claim 5 further comprising the steps of:

encoding each of the words in a common format of the same length, and

transmitting the words of each sentence in an implicit order recognizable by a receiver.

9. The encoding method of claim 1 wherein each elementary image block comprises a rectangular unit of 8 pixels on 8 adjacent lines of each of 2 consecutive frames.