Image analysis – Image compression or coding – Transform coding
Reexamination Certificate
2002-01-07
2003-11-04
Boudreau, Leo (Department: 2621)
Image analysis
Image compression or coding
Transform coding
C382S233000, C382S246000
Reexamination Certificate
active
06643408
ABSTRACT:
BACKGROUND OF THE INVENTION
(1) Field of the Invention
The present invention relates to an apparatus and a method for an orthogonal transform, a quantization, and an entropy coding of images and a medium recording a program for these coding processes, and relates to an apparatus and a method for an inverse transform and the decoding of a compressed code sequence which is obtained after the orthogonal transform, quantization, and entropy coding, and a recoding medium for recording a program for these decoding processes.
(2) Description of the Related Art
In recent years, techniques for decoding images which have been coded and compressed with high efficiency have been extensively developed and used in various fields, such as computers, communications, and broadcasts. Joint Photographic Coding Experts Group (JPEG) is recommended as an international standard for the compression coding and expansion decoding of images and Moving Picture Experts Group 1 (MPEG 1) Video is recommended as an international standard for the decoding of movies.
With these methods, image information is divided into blocks, each of which is composed of a certain number of pixels. Each pixel in the blocks is subjected to the orthogonal transform, quantization, and entropy coding to generate codes (or sometimes referred to as Huffman codes). The generated codes are combined and construct a compression code sequence (or referred to simply as a code sequence). Codes in the compression code sequence are subjected to the entropy decoding, dequantization, and inverse orthogonal transform to generate an original image.
It should be noted here that, in the following description, the process for subjecting each pixel to the orthogonal transform, quantization, and entropy coding is referred to as a compression coding (or referred to simply as a coding) to be distinguished from the entropy coding. Also, the process for subjecting codes to the entropy decoding, dequantization, and inverse transform is referred to as an expansion decoding (or referred to simply as a decoding) to be distinguished from the entropy decoding.
1. Operation for Compression Coding and Decoding
The following is a description of the operation for the compression coding and the expansion decoding which is base on the baseline method which is essential to an image coding apparatus and an image decoding apparatus supporting Discrete Cosine Transform (DCT) of JPEG. The following description uses an example where each of an image coding apparatus and an image decoding apparatus includes a general-purpose microcomputer.
1.1 Procedure for Compression Coding
(1) Outline of Operation Procedure for Compression Coding
The following is a description of the operation procedure for the compression coding, with reference to the drawings.
FIG. 1
is a flowchart showing the outline of the operation procedure for the compression coding which is based on the JPEG baseline method.
Initialization for coding control is performed and one or, more marker codes prescribed by JPEG are outputted (Step S
1901
). A rectangular area called a block composed of 8×8 pixels is cut out of an image to be compressed and coded (Step S
1902
), each element in the cut-out block is subjected to the discrete cosine transform, namely one type of the orthogonal transform, DCT coefficients of 8 lines×8 columns are outputted (Step S
1903
), the outputted DCT coefficients are divided by their corresponding elements in a quantization table, and 64 quantized DCT coefficients are outputted (Step S
1904
). The process for dividing each DCT coefficient by its corresponding element in the quantization table is called a quantization. The quantized DCT coefficients are scanned in the order called a zigzag scan and are subjected to the entropy coding, namely one type of the transform coding, to output a compression code sequence in which an image is compressed (Step S
1905
). Whether this process is repeated for all blocks of the image to be compressed and coded is confirmed (Step S
1906
). If all blocks of the image have been subjected to this process, the compression coding process is completed; if not, the operation returns to Step S
1902
and this process is commenced for the next block.
(2) Operation Procedure for Quantization
The following is a description of the operation procedure in the quantization step shown in
FIG. 1
as Step S
1904
, with reference to the flowchart shown in FIG.
2
.
The loop counter “i” for controlling the repetition of the quantization process is initialized to be set to the value “0” (Step S
1911
). A DCT coefficient S(i) and an element Q(i) in the quantization table, which each correspond to the value of the loop counter “i”, are loaded from the memory (Steps S
1912
and S
1913
), S(i) is divided by Q(i), the division result is rounded to be normalized so that a quantized DCT coefficient Sq(i) is obtained (Step S
1914
), and the quantized DCT coefficient Sq(i) is stored in a memory (Step S
1915
). The loop counter “i” is incremented by one to quantize the next DCT coefficient (Step S
1916
), and whether all DCT coefficients in one block have been subjected to this process is confirmed using the value of the loop counter “i” (Step,S
1917
). If all DCT coefficients in one block have been subjected to this process, this quantization process is completed; if not, the operation returns to Step S
1912
and remaining DCT coefficients are subjected to this quantization process.
(3) Operation Procedure for Entropy Coding
The following is a description of the operation procedure in the entropy coding step shown in
FIG. 1
as Step S
1905
, with reference to the flowchart in FIG.
3
.
A Sq(
0
), namely the quantized DCT coefficient of a DC component, is loaded from the memory and the DC component Sq(
0
) is subjected to the entropy coding (Steps S
1921
and S
1922
). A loop counter “i” for controlling the repetition of the entropy coding process on the quantized DCT coefficients of AC components and a counter “nnnn” for counting each zero coefficient run are initialized to be set to the values “1” and “0,” respectively (Step S
1923
). A scan order “z” corresponding to the loop counter “i” is obtained using a zigzag function “Zig( )” for generating a scan order used to perform a zigzag scan on quantized DCT coefficients (Step S
1924
). Using the scan order “z,” a quantized DCT coefficient Sq(z) of an AC component is loaded from the memory (Step S
1925
) and whether the value of this loaded quantized DCT coefficient Sq(z) is zero is judged (Step S
1926
). If the value of the quantized DCT coefficient Sq(z) is zero, the zero coefficient run counter “nnnn” is incremented by one (Step S
1931
); if not, the value of the zero coefficient run counter “nnnn” and the quantized DCT coefficient Sq(z) are combined and subjected to the entropy coding (Step S
1927
), and the zero coefficient run counter “nnnn” is initialized to be set to the value “0” (Step S
1928
). The loop counter “i” is incremented by one (Step S
1929
) and whether this process has been performed on one block is judged using the value of the loop counter “i” (Step S
1930
). If this process has been performed on one block, the entropy coding is completed; if not, the operation returns to Step S
1924
to continue this entropy coding.
1.2 Procedure for Expansion Decoding
To use compressed and coded images, these images need to be subjected to the expansion decoding. The following is a description of the procedure for the expansion decoding, with reference to the drawings.
(1) Outline of Operation Procedure for Expansion Decoding
FIG. 4
is a flowchart showing the outline of the expansion decoding which is based on the JPEG baseline method.
Initialization for control of expansion decoding is performed and each marker code included in a compression code sequence is analyzed (Step S
1941
). The compression code sequence is subjected to the entropy decoding, namely one type of the transform coding, and 64 quantized DCT coefficients corresponding to one block are obtained (Step S
1942
). The obtained quantized DCT c
Boudreau Leo
Matsushita Electric - Industrial Co., Ltd.
Sherali Ishrat
LandOfFree
Image coding and decoding apparatus, method of image coding... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image coding and decoding apparatus, method of image coding..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image coding and decoding apparatus, method of image coding... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3141221