Image analysis – Image compression or coding
Reexamination Certificate
1998-02-18
2001-06-19
Couso, Jose L. (Department: 2621)
Image analysis
Image compression or coding
C382S260000
Reexamination Certificate
active
06249610
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a picture coding apparatus and a method for coding digital picture signals and a picture decoding apparatus and a method for decoding coded digital picture signals.
BACKGROUND ART
Recently, as digital signal processing technology makes progress, coding apparatus for still pictures in digital still cameras, photo videotex and so on and coding apparatus for moving pictures utilized in video telephones, videoconferencing system, video CD system, DVD system, digital VTR(image processing system) and so on have been developed. For instance, coding apparatus utilizing orthogonal transform described in “Multi-dimensional Signal Processing for TV Pictures” written by Takahiko FUKINUKE (From page 213 to page 291, Chapter VII:High-efficient Coding, published Nov. 15, 1988 by Nikkan Kogyo Papers) and coding apparatus utilizing motion compensated interframe prediction are known.
In these apparatus, a digital picture discreted in M pixel in horizontal direction and N line in vertical direction are divided in blocks having m pixels in horizontal direction and n lines in vertical direction, and, in each block, high-efficient coding by motion compensated prediction processing, orthogonal transform processing and quantization processing orthogonal transform coefficient is carried out.
A conventional picture coding apparatus and a conventional picture decoding apparatus utilizing discrete cosine transform(abbreviated as DCT in the following) are explained below by using FIG.
9
.
In
FIG. 9A
,
1
is an input terminal to which digital picture signals is input,
2
is an output terminal for outputting coded input picture signals,
11
is DCT processing section for processing two-dimensional DCT to each block of digital picture divided in blocks,
12
is quantizing section for quantizing DCT coefficients obtained by DCT processing section
11
and
13
is DCT coefficient coding section for transgenerating quantized DCT coefficients into code word.
And, in
FIG. 9B
,
3
is an input terminal to which code word is input,
4
is an output terminal for outputting decoded digital pictures,
22
is DCT coefficient decoding section for reproducing quantized DCT coefficients from code word,
23
dequantizing section to obtain non-quantized DCT coefficients from quantized DCT coefficients and
24
is inverse two-dimensional DCT processing section for reproducing a picture from DCT coefficients obtained by dequantizing section
23
.
Each operation of the picture coding apparatus and the picture decoding apparatus constituted as described above is explained.
In the picture coding apparatus, an input picture is transformed into digital signals by an analog-digital converting circuit which is not shown and the input picture transformed into digital signals is divided in blocks in m pixels in horizontal direction and n lines in vertical directions. Each block of the input picture divided in block unit is input from input terminal
1
as digital picture signal
101
. DCT processing section
11
acquires digital signal
101
comprised of one block as one unit and processes two-dimensional DCT to output DCT coefficient
102
. Quantizing section
12
quantizes DCT coefficient
102
received from DCT processing section
11
and outputs DCT coefficient
103
. Coding section
13
outputs code word
104
corresponding to quantized DCT coefficient
103
received from quantizing section
12
.
In the picture decoding apparatus, code word
104
provided from input terminal
3
is transformed into quantized DCT coefficient
103
by DCT coefficient decoding section
22
. Dequantizing section
23
transforms quantized DCT coefficient
103
input from DCT coefficient decoding section
22
into DCT coefficient
108
.
Inverse two-dimensional DCT processing section
24
processes inverse two-dimensional DCT to DCT coefficient
108
to obtain reproduced picture
109
.
By the way, block dividing digital pictures then coding a digital picture by quantizing orthogonal transform and orthogonal transform coefficient in block unit as described above results in the difference between the signal value before coding processing and the decoded signal value in a block, which is caused by error in DCT/inverse DCT and quantizing error. That brings signal non-continuity in neighbor blocks, then noise called block distortion generates.
As measures to prevent this noise, a method for giving mask processing to orthogonal transform coefficient in picture coding apparatus (for instance, Separate volume 1-202 of Collections of National Conference Lecture Theses in Information/System Division of Electronic Communication Society in 1985) and a method for restoring a decode picture by using a post filter in a picture decoding apparatus (for instance, Separate volume 7 D-307 of Collections of Spring National Conference Lecture Theses in Electronic Information Communication Society in 1990) are proposed.
However, these methods can not decide which makes the non-continuity in block boundary, error caused by coding-decoding processing or edge component which the original input picture have, then it is impossible to completely prevent noise generation and to collect the reproduced picture.
DISCLOSURE OF INVENTION
The present invention is based on the facts described above.
The first purpose of the present invention is to provide an apparatus and a method for coding a picture which generate filter information used to appropriately correct non-continuity on a block boundary caused by error in coding/decoding processing.
And the second purpose of the present invention is to provide an apparatus and a method for decoding a picture which acquire the filter information used to appropriately correct non-continuity on a block boundary from input data, correct a reproduced picture and obtain high quality reproduced picture.
In the present invention, in the picture coding apparatus, a coded picture is reproduced once, noise component in the reproduced pictures is detected by comparing the reproduced picture and the input picture, a correction filter for the reproduced picture is selected corresponding to the characteristics of the noise then information on the selected correction filter for the reproduced picture is provided into the picture decoding apparatus for reproducing a coded picture.
Namely, in the coding apparatus, a direction and scale of non-continuity of signal value on a block boundary in the reproduced picture and whether the edge of the input picture causes the generation of the non-continuity or not are judged, and, based on the judged results, characteristics of filter and timing of turning on or off the filter are selected. Then, in the decoding apparatus, appropriate filter processing is carried out to the only part which requires correction by filter. These steps bring improvement of picture quality in a whole picture.
In the present invention, in the picture coding apparatus, by preparing a picture reproducing section at the latter part in the picture coding section and comparing the input picture and the reproduced picture, noise existing in the reproduced picture is detected, an appropriate filter according to the characteristics of the detected noise is selected then multiplexed information concerning the picture and filter is transmitted to establish the purpose described above.
And in the picture decoding apparatus, the information concerning the picture and the information concerning the filter are extracted from the transmitted information by inverse multiplex processing it, the picture is reproduced from the information concerning the picture then correction processing to the reproduced picture is carried out based on the information concerning the filter to establish the purpose described above.
A noise generation and characteristics of the noise in the reproduced picture obtained by coding processing then decoding processing are expected to be measured by calculating finite difference value between signal values of input picture and output picture. For instance, in the case where the fin
Matsumoto Taisuke
Tanaka Akiyoshi
Couso Jose L.
Dang Duy M.
Greenblum & Bernstein P.L.C.
Matsushita Electric - Industrial Co., Ltd.
LandOfFree
Apparatus and method for coding a picture and apparatus and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for coding a picture and apparatus and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for coding a picture and apparatus and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2466715