Image compressing device and image compressing method

Patent metadata

Field	Value
Publication number	US-10129543-B2
Application number	US-201715461621-A
Country	US
Kind code	B2
Filing date	Mar 17, 2017
Priority date	Mar 28, 2016
Publication date	Nov 13, 2018
Grant date	Nov 13, 2018

Abstract

A prefetched-image memory stores a prefetched image of a prefetched area prefetched from a reference image memory that stores a reference image. A processor performs a motion search using a reduced image obtained by reducing an encoding target image and generates a search result for an encoding target block included in the encoding target image. The processor determines, based on the search result, that an area situated outside the prefetched area in the reference image is to be searched when the processor has predicted from the search result that a matching of the encoding target block and the prefetched image will be unsuccessful. Next, the processor performs a matching of the encoding target block and an image of a matching target area including the outside area in the reference image and generates an inter-prediction result. Then, the processor encodes the encoding target block based on the inter-prediction result.

First claim

What is claimed is:

1. An image compressing device comprising:
a prefetched-image memory that stores a prefetched image of a prefetched area prefetched from a reference image memory that stores a reference image; and
a processor that
performs a motion search using a reduced image obtained by reducing an encoding target image and generates a search result for an encoding target block included in the encoding target image,
determines, on the basis of the search result, that an area situated outside the prefetched area in the reference image is to be searched when the processor has predicted from the search result that a matching of the encoding target block and the prefetched image will be unsuccessful,
performs a matching of the encoding target block and an image of a matching target area including the outside area in the reference image when the processor has determined that the outside area is to be searched and generates an inter-prediction result, and
encodes the encoding target block on the basis of the inter-prediction result.

2. The image compressing device according to claim 1 , wherein
the processor determines the matching target area in the reference image on the basis of the search result, and predicts that the matching of the encoding target block and the prefetched image will be unsuccessful when the matching target area is not within the prefetched area.

3. The image compressing device according to claim 2 , wherein
the processor calculates an evaluation value of a number of accesses to be made to the reference image memory in order to read an image of the outside area, and determines whether the outside area is to be searched, on the basis of the evaluation value.

4. The image compressing device according to claim 3 , wherein
the processor generates a plurality of search results respectively corresponding to a plurality of blocks included in the encoding target image, calculates a plurality of evaluation values respectively corresponding to the plurality of blocks when the processor has predicted that a matching will be unsuccessful for the plurality of blocks, and determines whether the outside area is to be searched with respect to each of the plurality of blocks, on the basis of a result of comparing the plurality of evaluation values.

5. The image compressing device according to claim 4 , wherein
the processor weights each of a plurality of numbers of accesses respectively corresponding to the plurality of blocks by using a dissimilarity between each of the plurality of blocks respectively indicated by the plurality of search results and the reference image, so as to calculate the plurality of evaluation values.

6. The image compressing device according to claim 1 , wherein
the reference image memory further stores another reference image that is different from the reference image,
the prefetched-image memory further stores another prefetched image of a prefetched area prefetched from the another reference image, and
when the processor has predicted that a matching of the encoding target block and the another prefetched image will be successful, the processor decreases a probability that the outside area will be determined to be searched with respect to the encoding target block.

7. An image compressing method comprising:
performing, by an image compressing device, a motion search using a reduced image obtained by reducing an encoding target image and generating a search result for an encoding target block included in the encoding target image, the image compressing device storing a prefetched image of a prefetched area prefetched from a reference image memory that stores a reference image;
determining, by the image compressing device and on the basis of the search result, that an area situated outside the prefetched area in the reference image is to be searched when a matching of the encoding target block and the prefetched image has been predicted to be unsuccessful from the search result;
performing, by the image compressing device, a matching of the encoding target block and an image of a matching target area including the outside area in the reference image and generating an inter-prediction result; and
encoding, by the image compressing device, the encoding target block on the basis of the inter-prediction result.

8. The image compressing method according to claim 7 , further comprising:
determining the matching target area in the reference image on the basis of the search result; and
predicting that the matching of the encoding target block and the prefetched image will be unsuccessful when the matching target area is not within the prefetched area.

9. The image compressing device according to claim 8 , further comprising:
calculating an evaluation value of a number of accesses to be made to the reference image memory in order to read an image of the outside area; and
determining whether the outside area is to be searched, on the basis of the evaluation value.

10. The image compressing method according to claim 9 , wherein
the generating the search result generates a plurality of search results respectively corresponding to a plurality of blocks included in the encoding target image,
the calculating the evaluation value calculates a plurality of evaluation values respectively corresponding to the plurality of blocks when a matching has been predicted to be unsuccessful for the plurality of blocks, and
the determining whether the outside area is to be searched determines whether the outside area is to be searched with respect to each of the plurality of blocks, on the basis of a result of comparing the plurality of evaluation values.

11. A non-transitory computer-readable recording medium having stored therein an image compressing program causing a computer to execute a process, the computer storing a prefetched image of a prefetched area prefetched from a reference image memory that stores a reference image, the process comprising:
performing a motion search using a reduced image obtained by reducing an encoding target image and generating a search result for an encoding target block included in the encoding target image;
determining, on the basis of the search result, that an area situated outside the prefetched area in the reference image is to be searched when a matching of the encoding target block and the prefetched image has been predicted to be unsuccessful from the search result;
performing a matching of the encoding target block and an image of a matching target area including the outside area in the reference image and generating an inter-prediction result; and
encoding the encoding target block on the basis of the inter-prediction result.

12. The non-transitory computer-readable recording medium according to claim 11 , further comprising:
determining the matching target area in the reference image on the basis of the search result; and
predicting that the matching of the encoding target block and the prefetched image will be unsuccessful when the matching target area is not within the prefetched area.

13. The non-transitory computer-readable recording medium according to claim 12 , further comprising:
calculating an evaluation value of a number of accesses to be made to the reference image memory in order to read an image of the outside area; and
determining whether the outside area is to be searched, on the basis of the evaluation value.

14. The non-transitory computer-readable recording medium according to claim 13 , wherein
the generating the search result generates a plurality of search results respectively corresponding to a plurality of blocks includ

Assignees

• Fujitsu Ltd

Inventors

• Tanaka Ayana
• Misuda Yasuo
• Imajo Chikara

Classifications

• H04N19/105

  Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction · CPC title

• H04N19/53

  Multi-resolution motion estimation; Hierarchical motion estimation · CPC title

• H04N19/51

  Motion estimation or motion compensation · CPC title

• H04N19/176

  the region being a block, e.g. a macroblock · CPC title

• H04N19/159

  Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction · CPC title