Methods and systems for image compression

Patent metadata

Field	Value
Publication number	US-11019365-B2
Application number	US-201615747589-A
Country	US
Kind code	B2
Filing date	Jul 29, 2016
Priority date	Jul 30, 2015
Publication date	May 25, 2021
Grant date	May 25, 2021

Abstract

A method relating to image compression includes acquiring image data of an image, wherein the image includes a matrix of pixels, and the image data include pixel values of the matrix of pixels. The method also includes generating, based on the image data, a first-level data set including pixel values of a portion of the matrix of pixels. The method further includes determining at least one coding mode for the first-level data set, encoding the first-level data set into a code stream based on the at least one coding mode, and packaging the code stream into a modified code stream.

First claim

What is claimed is:

1. A method implemented on at least one machine, each of which has at least one processor and storage device for encoding an image, the method comprising:
acquiring image data of an image, the image being a color filter array image, the image including a matrix of pixels, the image data including pixel values of the matrix of pixels, each pixel of the matrix of pixels corresponding to a chroma component;
encoding the image row by row, wherein the encoding of each row of the matrix of pixels includes:
generating, based on a color filter array pattern of the each row of the matrix of pixels and chroma components of the pixels in the each row, a first-level data set corresponding to the each row;
determining a first second-level data set based on the first-level data set, the first second-level data set representing coefficients of different direct component values corresponding to pixels of the each row;
determining a second second-level data set based on the first second-level data set, an inverse quantization process, and an inverse 2D DCT process, the second second-level data set representing chroma component related residual values corresponding to the each row;
determining a first third-level data set based on the first second-level data set;
determining a second third-level data set based on the second second-level data set;
determining a first coding mode for the first third-level data set by predicting lengths of code streams, with a first set of candidate coding modes, for the first third-level data set;
determining a second coding mode for the second third-level data set by predicting lengths of code streams, with a second set of candidate coding modes, for the second third-level data set;
encoding the first third-level data set and the second third-level data set into a code stream based on the first coding mode and the second coding mode; and
packaging the code stream into a modified code stream corresponding to the each row.

2. The method of claim 1 , the determining a first second-level data set based on the first-level data set comprising:
generating at least one chroma component data set each of which corresponding to a same chroma component by extracting, based on the color filter array pattern of the each row, pixels corresponding to the same chroma component;
generating a plurality of transformation coefficients by performing a two-dimensional discrete cosine transform (2D DCT) process and a quantization process on each of the at least one chroma component data set; and
generating, based on the plurality of transformation coefficients, the first second-level data set.

3. The method of claim 2 , generating at least one chroma component data set comprising:
generating a first chroma component data set by extracting pixels at odd positions of the each row; and
generating a second chroma component data set by extracting pixels at even positions of the each row.

4. The method of claim 1 , the determining the second second-level data set based on the first second-level data set comprising:
generating a reconstruction data set by performing the inverse quantization process and the inverse 2D DCT process on the first second-level data set;
calculating the chroma component related residual values based on a difference between the pixels of the each row and the reconstruction data set; and
generating the second second-level data set based on the chroma component related residual values corresponding to the each row.

5. The method of claim 1 , further comprising:
partitioning the first second-level data set into N1 third-level data sets that form the first third-level data set, N1 being a second integer, the second integer being greater than or equal to 1; and
partitioning the second second-level data set into N2 third-level data sets that form the second third-level data set, N2 being a third integer, the third integer being greater than or equal to 1.

6. The method of claim 5 , partitioning the first second-level data set into N1 third-level data sets comprising:
partitioning each pixel value of the first second-level data into N1 portions, each portion of the N1 portions including one or more bits; and
generating each third-level data set by extracting a portion of the N1 portions of the each pixel value.

7. The method of claim 5 , partitioning the second second-level data set into N2 third-level data sets comprising:
partitioning each pixel value of the second second-level data into N2 portions, each portion of the N2 portions including one or more bits; and
generating each third-level data set by extracting a portion of the N2 portions of the each pixel value.

8. The method of claim 5 , wherein the first second-level data set includes one or more component values of transformation coefficients, and the packaging the code stream into a modified code stream corresponding to the each row comprises:
generating a header, the header including a coding mode of each of the N1 third-level data sets, a length of a code stream of each of the N1 third-level data sets, a coding mode of each of the N2 third-level data sets, a length of a code stream of each of the N2 third-level data sets; and
packaging the header into the modified code stream, the modified code stream including the one or more component values of transformation coefficients.

9. The method of claim 5 , wherein the first second-level data set includes one or more component values of transformation coefficients, and the partitioning the first second-level data set into N1 third-level data sets comprises:
extracting, from the first second-level data set, data excluding the one or more component values of transformation coefficients;
mapping the extracted data into non-negative integers;
partitioning each of the non-negative integers into N1 portions, each of the N1 portions including one or more bits; and
generating each third-level data set by extracting a portion of the N1 portions of the each non-negative integer.

10. The method of claim 1 , determining the first coding mode for the first third-level data set by predicting lengths of code streams, with the first set of candidate coding modes, for the first third-level data set comprising:
for each first third-level data set,
predicting lengths of code streams of the first third-level data set with at least one of Huffman coding, run-length coding, fixed-length coding, or order-k exponential Golomb coding;
comparing the predicted lengths; and
selecting the first coding mode based on a result of the comparison.

11. The method of claim 10 , determining the first coding mode for the first third-level data set by predicting lengths of code streams, with the first set of candidate coding modes, for the first third-level data set further comprising:
for the each first third-level data set,
calculating an original length of the code stream of the first third-level data set without coding;
comparing the predicted lengths with the original length; and
determining whether to encode the first third-level data set based on a result of the comparison.

12. The method of claim 11 , determining whether to encode the first third-level data set comprising:
determine to encode the first third-level data set using the coding mode with the shortest length of the predicted lengths, if the predicted lengths are less than the original length; and
determine not to encode the first third-level data set, if any length of predicted lengths is not less than the original length.

13. The method of claim 1 , wherein the each pixel of the matrix of pixels corresponds to only one chroma compon

Assignees

• Zhejiang Dahua Technology Co

Inventors

• Lu Jingang
• Chen Weigang
• Yao Weizhong
• Zhang Xingming
• Yang Yinchang

Classifications

• H04N19/119 Primary

  Adaptive subdivision aspects, e.g. subdivision of a picture into rectangular or non-rectangular coding blocks · CPC title

• H04N19/182

  the unit being a pixel · CPC title

• H04N19/174

  the region being a slice, e.g. a line of blocks or a group of blocks · CPC title

• H04N19/186

  the unit being a colour or a chrominance component · CPC title

• H04N19/30

  using hierarchical techniques, e.g. scalability (H04N19/63 takes precedence) · CPC title