Motion video compression system with multiresolution features

• US 5,414,469 A
• Filed: 10/31/1991
• Issued: 05/09/1995
• Est. Priority Date: 10/31/1991
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Viscito, Eric, Ramm, Dov, Gonzales, Cesar A.
• Primary Examiner(s): Herndon, Heather R.
• Assistant Examiner(s): BURRASTON, NEIL
• Application Number:
  US07/786,233
• Time in Patent Office: 1,286 Days
• Field of Search: 395/152, 382/56, 358/136, 358/433, 364/715.02, 348/407, 348/408, 348/409, 348/415, 348/416
• US Class Current: 375/240.18
• CPC Class Codes: H04N 19/53 Multi-resolution motion estimation; Hierarchical motion estimation
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/63 using sub-band based transform, e.g. wavelets

Abstract

A system and method for processing a stream of video image data so as to create a video representation that multiplexes data corresponding to resolution or bitstream scales. This representation is such that the identity of the basic MacroBlock (MB) structure of the MPEG-1 ISO standard is preserved across all resolution and bitstream scales, e.g. by scaling across four levels of resolution. A MacroBlock is associated with a series of attributes which contribute to the amount of overhead data incorporated in an MPEG-1 compressed data stream, so that by preserving the MacroBlock identify across multiple resolutions and bitstream scales, these scales can share this overhead, thus requiring it to be included only once in the data stream. Preserving the MacroBlock identify also simplifies significantly the derivation of motion estimation vector data for all resolution scales other than the highest resolution. Essentially, the motion vector data corresponding to any resolution scales can be derived from the highest resolution motion vector by appropriately scaling it down. Alternatively, the full resolution motion vectors can be derived by appropriate scale up of lower resolution motion vectors. Further, the methodology for coding a MacroBlock is also preserved.

Claims

1. A method of decoding a moving image at a target spatial resolution scale from compressed video information containing data for a hierarchy of spatial resolution scales, comprising the steps of:

(a) decompressing macroblock attributes and transform coefficient data from a base spatial resolution scale;

(b) decompressing refinements to the macroblock attributes and additional transform coefficient data for the next higher spatial resolution scale in the hierarchy;

(c) repeating step (b) for progressively higher spatial resolution scales in the hierarchy until the target spatial resolution scale is reached; and

,

(d) combining the transform coefficient data which have been decompressed to produce a transform representation of the moving image.

2. A method as in claim 1 wherein said macroblock attributes are appropriately modifiable to conform to the spatial resolution scale of said target spatial resolution scale.

3. A method as in claim 1 further comprising the step of omitting at least a portion of the compressed under information in the macroblock when found in error.

4. A method of producing a compressed video data representation of a moving image, the representation including a hierarchy of picture spatial resolution scales, comprising the steps of:

providing compressed video data representative of the moving image for a base spatial resolution scale, the data comprising macroblock attributes common to the hierarchy and associated base resolution transform data;

providing additional data for at least one higher spatial resolution scale, the additional data including associated higher resolution transform data;

wherein the transform data for each spatial resolution scale in the hierarchy is used to predictably generate the higher resolution transform data for the next higher spatial resolution scale in the hierarchy.

5. The method of claim 4 wherein the additional data for higher spatial resolution scales includes refinements of the macroblock attributes at the base resolution scale.

6. A method of producing a compressed video data representation of a moving image, the representation including a hierarchy of picture quality scales at the same spatial resolution, comprising the steps of:

providing compressed video data representative of the moving image for a base picture quality scale, the data comprising macroblock attributes and associated base bitstream transform data; and

,

providing additional data for at least one higher picture quality scale, the additional data including refinement data to increase the precision of the transform data;

wherein the transform data for each picture quality scale in the hierarchy is used to predictably generate higher precision transform data for the next picture quality scale in the hierarchy.

7. The method of claim 6 comprising the further steps of:

providing base data for a base resolution scale, the base data comprising the macroblock attributes and associated base resolution transform data;

providing higher resolution scale data for at least one additional higher resolution scale, the higher resolution scale data including refinements of the macroblock attributes at the base resolution scale and associated higher resolution transform data;

wherein the base resolution transform data is used to predictably generate the higher resolution transform data.

8. The method of claim 6 wherein the additional data for higher precision picture quality scales includes refines of the macroblock attributes at the base picture quality scale.

9. A method of producing a compressed video data representation of a moving image, the representation including a hierarchy of spatial resolution scales, comprising the steps of:

providing compressed video data representative of the moving image for a base spatial resolution scale, the compressed video data comprising macroblock attributes common to the hierarchy and associated base resolution transform data; and

,

starting with the base spatial resolution scale, using the transform data for each of the spatial resolution scales to predictably generate higher resolution transform data for a next higher spatial resolution scale until the transform data has been generated for each spatial resolution scale in the hierarchy.

10. A method of decoding a moving image at a target quality level scale from compressed video information containing data for a hierarchy of quality level scales, comprising the steps of:

(a) decompressing macroblock attributes and transform coefficient data from a base quality level scale;

(b) decompressing refinements to the macroblock attributes and additional transform coefficient data for the next higher quality level scale in the hierarchy;

(c) repeating step (b) for progressively higher quality resolution scales in the hierarchy until the target quality resolution scale is reached; and

(d) combining the transform coefficient data which have been decompressed to produce a transform representation of the moving image.