Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system

• US 6,404,815 B1
• Filed: 01/19/1999
• Issued: 06/11/2002
• Est. Priority Date: 03/17/1997
• Status: Expired due to Term

• Current Assignee: Mitsubishi Denki Kabushiki Kaisha (Mitsubishi Electric Corporation)
• Original Assignee: Mitsubishi Denki Kabushiki Kaisha (Mitsubishi Electric Corporation)
• Inventors: Nishikawa, Hirofumi, Asai, Kohtaro, Hasegawa, Yuri, Sekiguchi, Shunichi, Isu, Yoshimi, Murakami, Tokumichi, Kuroda, Shinichi
• Primary Examiner(s): Le, Vu
• Application Number:
  US09/180,188
• Time in Patent Office: 1,239 Days
• Field of Search: 375/240.16, 375/240.17, 375/240.24, 375/240.18, 375/240.29, 382/236, 382/238, 382/243, 382/248, 382/109, 382/111, 358/261.2, 358/428, 358/430, 348/416.1, 348/413.1, 348/415.1, 348/699, 348/700
• US Class Current: 375/240.16
• CPC Class Codes: H04N 19/523 with sub-pixel accuracy
• CPC Class Codes: H04N 19/53 Multi-resolution motion estimation; Hierarchical motion estimation
• CPC Class Codes: H04N 19/537 Motion estimation other than block-based
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

An efficient video encoder and a compression encoding of an input picture includes motion compensated prediction means for performing interframe motion detection. The motion compensated prediction means has a motion detector and a motion compensator. The motion detector includes an image transforming unit that transforms a reference picture portion into a picture portion having a predefined shape. The motion compensator includes a pel determinator for determining integer pels of real sample points of the transformed reference picture portion that corresponds to an integer pel of the input picture block according to the motion parameters supplied from the motion detector as a result of the comparison so as to supply prediction picture data. The video encoder and a decoder having a motion compensator that performs operations similar to the video encoder are combined to form a video encoding and decoding system.

Claims

1. A video encoder for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detector for detecting at least one motion parameter providing minimum prediction error, including an image transformation unit for transforming a reference image portion into a transformed image portion using a given transformation, and for matching the transformed reference image portion with an input image portion;

a motion compensator for outputting a motion compensated prediction image, including a corresponding pel determinator for determining pels for transforming the reference image portion based on the motion parameters;

a multiplexer for multiplexing said motion parameters together with the information which indicates a given transformation to the encoded bitstream; and

wherein the image transformation unit transforms a reference image portion comprised of half-pels as well as integer pels into the transformed image portion.

9. The video encoder according to claim 1, wherein the reference image portion is one of the following candidates:

1) a portion which is contained in the frame which is forward in time with respect to the frame which contains the input image portion,

     2) a portion which is contained in the frame which is backward in time with respect to the frame which contains the input image portion, and

     3) a portion which can be obtained by averaging

     1) and

     2).

2. A video encoder for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detector for detecting at least one motion parameter providing minimum prediction error, including an image transformation unit for transforming a reference image portion into a transformed image portion using a given transformation, and for matching the transformed reference image portion with an input image portion;

a motion compensator for outputting a motion compensated prediction image, including a corresponding pel determinator for determining pels for transforming the reference image portion based on the motion parameters;

a multiplexer for multiplexing said motion parameters together with the information which indicates a given transformation to the encoded bitstream; and

a preprocessor for separating an input video into video objects each of which is an coding unit, the video encoder dividing the video object into blocks, detecting motion parameters, and compensating the motion of a divided video object.

3. A video encoder for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detector for detecting at least one motion parameter providing minimum prediction error, including an image transformation unit for transforming a reference image portion into a transformed image portion using a given transformation, and for matching the transformed reference image portion with an input image portion;

a motion compensator for outputting a motion compensated prediction image, including a corresponding pel determinator for determining pels for transforming the reference image portion based on the motion parameters;

a multiplexer for multiplexing said motion parameters together with the information which indicates a given transformation to the encoded bitstream; and

wherein the image transformation unit transforms the reference image portion comprised of half-pels as well as integer pels by indicating addresses of one of an image portion scaled down at a specific ratio and an image portion scaled up at a specific ratio, and matches the transformed reference image portion with the input image portion.

4. A video encoder for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detector for detecting at least one motion parameter providing minimum prediction error, including an image transformation unit for transforming a reference image portion into a transformed image portion using a given transformation, and for matching the transformed reference image portion with an input image portion;

a motion compensator for outputting a motion compensated prediction image, including a corresponding pel determinator for determining pels for transforming the reference image portion based on the motion parameters;

a multiplexer for multiplexing said motion parameters together with the information which indicates a given transformation to the encoded bitstream; and

wherein the image transformation unit transforms a reference image portion comprised of half-pels as well as integer pels by indicating addresses of an image portion rotated by a predefined angle.

5. The video encoder according to claim 4, wherein the image transformation unit indicates addresses of an image portion rotated by one of angles of ±

45°

, ±

90°

, ±

135°

, ±

180°

.

6. A video encoder for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detector for detecting a final motion vector providing minimum prediction error, the motion detector including;

a block matching unit for matching the reference image portion, which has been translated, with the input image portion to detect a first motion vector and produce a first minimum prediction error signal;

a transformed block matching unit for transforming a reference image portion consisting of half-pels as well as integer pels by indicating addresses of the half or integer pels into a transformed image portion having a predefined shape within a search range for a second motion vector around the location indicated by the first motion vector, and for matching the transformed reference image portion with an input image portion to detect a second motion vector and produce a second minimum prediction error signal;

a motion compensated prediction mode determinator for determining a final motion vector which provides the minimum prediction error based on the first and the second minimum prediction error signals;

a motion compensator for outputting a motion compensated prediction image, including a corresponding pel determinator for determining pels for transforming the reference image portion based on the motion vector; and

a multiplexer for multiplexing the final motion vector together with the information which indicates the predefined shape to the encoded bitstream.

7. A video encoder for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detector for detecting at least one motion parameter providing minimum prediction error, including an image transformation unit for transforming a reference image portion into a transformed image portion using a given transformation, and for matching the transformed reference image portion with an input image portion;

a motion compensator for outputting a motion compensated prediction image, including a corresponding pel determinator for determining pels for transforming the reference image portion based on the motion parameters;

a multiplexer for multiplexing said motion parameters together with the information which indicates a given transformation to the encoded bitstream; and

wherein the image transformation unit has a transformation pattern table for specifying addresses for transforming the reference image portion, and transforms the reference image portion by accessing the transformation pattern table.

8. A video encoder for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detector for detecting at least one motion parameter providing minimum prediction error, including an image transformation unit for transforming a reference image portion into a transformed image portion using a given transformation, and for matching the transformed reference image portion with an input image portion;

a motion compensator for outputting a motion compensated prediction image, including a corresponding pel determinator for determining pels for transforming the reference image portion based on the motion parameters;

a multiplexer for multiplexing said motion parameters together with the information which indicates a given transformation to the encoded bitstream; and

wherein the image transformation unit selectively filters specific pels of the reference image portion extracted for matching, and matches the transformed reference image portion with the input image portion.

10. A video decoder for decoding an encoded bitstream of video data, comprising:

a motion compensation unit including an image transformation unit for transforming a reference image portion into a transformed image portion using the transformation specified by at least one motion parameter decoded from an input bitstream;

an image reconstruction unit for reconstructing a decoded image portion using a decoded data and said transformed image portion; and

wherein the at least one motion parameter consists of a motion vector and a transformation pattern information, and the motion compensation unit includes a mechanism for determining pels of the corresponding reference image portion comprised of half-pels as well as integer-pels by transforming the corresponding reference image portion into one of an image portion scaled up, an image portion scaled down, and an image portion rotated according to the motion parameters.

11. A video decoder for decoding an encoded bitstream of video data, comprising:

a motion compensation unit including an image transformation unit for transforming a reference image portion into a transformed image portion using the transformation specified by at least one motion parameter decoded from an input bitstream;

an image reconstruction unit for reconstructing a decoded image portion using a decoded data and said transformed image portion; and

wherein the at least one motion parameter consists of plural motion vectors and a transformation pattern information, and the image transformation unit transforms the reference image portion comprised of half-pels as well as integer pels into the transformed image portion by indicating addresses calculated with plural motion vectors according to the transformation pattern information.

12. A video decoder for decoding an encoded bitstream of video data, comprising:

a motion compensation unit including an image transformation unit for transforming a reference image portion into a transformed image portion using the transformation specified by at least one motion parameter decoded from an input bitstream;

an image reconstruction unit for reconstructing a decoded image portion using a decoded data and said transformed image portion; and

wherein the at least one motion parameter consists of plural motion vectors, a transformation pattern information and a round-up information, and the image transformation unit transforms the reference image portion comprised of half-pels as well as integer pels into the transformed image portion by indicating addresses according to the following process;

1) calculate addresses with plural motion vectors according to the transformation pattern information,

2) round up calculated addresses with the round-up information.

13. A video decoder for decoding an encoded bitstream of video data, comprising:

a motion compensation unit including an image transformation unit for transforming a reference image portion into a transformed image portion using the transformation specified by at least one motion parameter decoded from an input bitstream;

an image reconstruction unit for reconstructing a decoded image portion using a decoded data and said transformed image portion; and

wherein the number of pels of the width and the height of the reference image portion is selected from a power of 2.

14. A video encoding method for performing motion compensated prediction encoding of block-divided input image, comprising:

a motion detecting step for detecting motion parameters providing minimum prediction error, including an image transformation step for transforming a reference image portion consisting of half-pels as well as integer pels into a transformed image portion using a given transformation, and for matching the transformed reference image portion with an input image portion;

a motion compensating step for outputting a motion compensated prediction image, including a corresponding pel determining step for determining pels for transforming the reference image portion based on the motion vector; and

a multiplexing step for multiplexing said motion parameters together with the information which indicates a given transformation to the encoded bitstream.

15. The video encoding method according to claim 14, wherein the image transformation step has a transformation pattern table for specifying addresses for transforming the reference image portion, and transforms the reference image portion by accessing the transformation pattern table.