Image analysis – Image compression or coding – Predictive coding
Reexamination Certificate
1998-07-20
2001-01-09
Lee, Thomas D. (Department: 2724)
Image analysis
Image compression or coding
Predictive coding
C382S247000, C382S239000
Reexamination Certificate
active
06173078
ABSTRACT:
TECHNICAL FIELD
The present invention relates to image coding apparatuses and image coding methods, image decoding apparatuses and image decoding methods, and data recording media and, more particularly, to a coding process and a decoding process for performing recording or transmission of image signals with less bit number without degrading the image quality, and to a recording medium containing a program for realizing the coding process or the decoding process.
BACKGROUND ART
Conventional image coding processes are broadly divided into two of coding processes performed in block units, represented by MPEG2-based coding methods, and coding processes performed in pixel units, such as differential pulse code modulation (DPCM: Differential Pulse Code Modulation).
The coding process in block units is a method in which a single image display region is divided into plural blocks and a coding process for an image signal that is input (hereinafter referred to as an input image signal) is performed block by block. In this case, the single image display region corresponds to a single display screen in the MPEG2-based coding process, and in an MPEG4-based coding process, it corresponds to a display region having a shape and a size corresponding to each object on a single display screen. Further, each of the above-mentioned blocks is a display region composed of a prescribed number of pixels within the single image display region, and a rectangle shape which is easily processed is used as the shape of the block in many cases.
As described above, in the coding method in which a coding process for an input image signal is performed block by block, the coding process for the input image signal corresponding to a single image display region is completed in each block. Therefore, there is the advantage that, even though a transmission error occurs when a coded image signal obtained by performing the coding process for the input image signal is transmitted, influence of the error can be converged in each block.
However, the block-by-block coding method has the following drawbacks.
Since the coding process for the input image signal is completed in each block in the block-by-block coding method, it is difficult to use an inter-block pixel correlation, that is, a correlation of pixel values which are present in different blocks, in the coding process.
Further, in a predictive coding method for an image signal, a pixel value of a coding target pixel being a target of coding (a coding target pixel value) is predicted with reference to pixel values of plural coded pixels which have previously been coded (coded pixel values), and the coding target pixel value is adaptively coded using the predicted pixel value. In this predictive coding method, however, when the coding process is performed block by block, the coded pixel values to be referred to when coding the coding target pixel value are limited to pixels within each block, so that the reference coded pixel values are small in number. Therefore, the accuracy of the predicted value of the coding target pixel is reduced, and the coding efficiency is not increased very much.
On the other hand, the coding method in pixel units is a method in which an input image signal is coded pixel by pixel, and in this coding method, it is possible to change the coding process for the input image signal pixel by pixel. Therefore, when this coding method includes a universal coding process, such as adaptive arithmetic coding in which code words are automatically updated pixel by pixel adaptively to the characteristics of the input image signal, an image signal having any characteristic can be coded with a significantly high coding efficiency.
However, since, on the decoding side, a coded image signal obtained by the pixel-by-pixel coding method including the universal coding process is subjected to a decoding process in which code words are updated in the same manner as on the coding side, when a transmission error occurs when the coded image signal is transmitted, the state in which the decoding process for the coded image signal cannot be carried out accurately because of the influence of the transmission error on the decoding side, continues long.
By the way, the block-by-block coding method and the pixel-by-pixel coding method can be combined, and in a coding method in which these coding methods are combined (hereinafter referring to this coding method as a combination coding method for explanation), code words can be adaptively changed for each pixel, and the influence of transmission error can be converged in each block, whereby a coding process with a high coding efficiency, such as adaptive arithmetic coding, can be performed while suppressing the influence of transmission error.
A description is now given of this combination coding method.
FIG.
13
(
a
) shows the state in which a single frame screen is divided into a plurality of rectangle blocks, and FIG.
13
(
b
) shows arrangement of pixels in blocks, especially in a coding target block being a target of coding and blocks in the vicinity of the coding target block. Needless to say, these pixels are arranged in matrix along horizontal scanning lines in the single frame screen.
In the figures, FG denotes a screen corresponding to a single frame, B
1
denotes a coded block in which a coding process for an image signal has already been performed, Bx denotes a coding target block being a target of coding, and B
0
denotes an uncoded block in which a coding process for an image signal has not been performed. When no distinction is made between blocks, blocks are denoted by B. BLu, BLs, BLh, and BLm denote upper, lower, left, and right boundaries of the coding target block in the single frame screen, respectively. A solid line circle shows a coded pixel whose pixel value has already been coded, and a dotted line circle shows an uncoded pixel whose pixel value has not been coded yet. Each block B is an image display region comprising 4×4 pixels, in the single frame screen FG.
FIG. 14
shows positional relationships between a coding target pixel Px to be coded and peripheral pixels P
0
~P
9
surrounding the coding target pixel, and the pixel values of these peripheral pixels P
0
~P
9
are referred to when the pixel value of the coding target pixel Px is predicted, so that these pixels are called reference pixels hereinafter.
The reference pixels P
8
and P
9
are pixels which are positioned in the same horizontal scanning line as the coding target pixel Px, and the reference pixels P
9
and P
8
are positioned one pixel and two pixels before the coding target pixel Px, respectively. The positions of the reference pixels P
5
and P
1
in the horizontal direction on the single frame screen FG correspond to the position of the coding target pixel Px, and the reference pixels P
5
and P
1
are positioned in a horizontal scanning line by one pixel and two pixels upper than the coding target pixel Px, respectively. Further, the reference pixels P
3
, P
4
, P
6
, and P
7
are pixels which are positioned in the same horizontal scanning line as the reference pixel P
5
, the reference pixels P
4
and P
3
are positioned one pixel and two pixels before the coding target pixel Px, respectively, and the reference pixels P
6
and P
7
are positioned one pixel and two pixels after the coding target pixel Px, respectively. Further, the reference pixels P
0
and P
2
are pixels which are positioned in the same horizontal scanning line as the reference pixel P
1
, the reference pixel P
0
is positioned one pixel before the reference pixel P
1
, and the reference pixel P
2
is positioned one pixel after the reference pixel P
1
.
In the combination coding method, initially, an image signal corresponding to the single frame screen FG is divided correspondingly to plural blocks B constituting the single frame screen as shown in FIG.
13
(
a
) and FIG.
13
(
b
), and a coding process for the divided image signal is performed block by block.
This block-by-block coding process is completed by performing a horizontal process in wh
Chen Wenpeng
Lee Thomas D.
Matsushita Electric - Industrial Co., Ltd.
Parkhurst & Wendel L.L.P.
LandOfFree
Image encoder, image encoding method, image decoder, image... 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 encoder, image encoding method, image decoder, image..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image encoder, image encoding method, image decoder, image... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2451455