Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-10-28
2001-10-16
Lee, Richard (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240180
Reexamination Certificate
active
06304606
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an image data coding method of coding the image data with a high efficiency, an apparatus for coding an image data, a restoring method of decoding the coded image and an apparatus for restoring the same and more particularly to an image data coding and decoding method of dividing the image data into a plurality of blocks to produce the coded data of all the images related to the respective blocks obtained by dividing the image data.
2. Description of the Related Art
In the field of a television conference/television telephone involving colored static images and colored moving image, the image data whose information capacity is much greater than typical numerical data are used, and more particularly, the data of the intermediate level of the image, or the colored image data requires high-speed and high quantity transmission. Therefore, it is necessary to process the coding of the values of the respective image elements with a high efficiency.
The adaptive discrete cosine transform method is known as an example of an image data coding method with a high efficiency.
This adaptive discrete cosine transform method (which is abbreviated as ADCT) divides the image into a plurality of blocks comprising, for example, 8×8 image elements, transforms the image signal of respective blocks into (two-dimensional DCT coefficients corresponding to the distributions of the space frequency by the two-dimensional discrete cosine transform), quantizes the two-dimensional DCT coefficient within a block obtained by the transform by using a threshold value suitable for vision, and then encodes the quantized coefficient by using a huffman table obtained statistically.
The conventional coding device for the static image using the ADCT encodes the respective images separately, even if the respective values between images do not change greatly because of a close interrelation between them, for example, in the case of a moving image which is considered as a group of static images.
FIG. 1
shows an example of a structure of the conventional coding apparatus for a static image.
A static image is divided into a plurality of blocks comprising 8×8 image elements, for example, and is inputted from the input terminal
11
and the respective image signals are sequentially restored in the input order to the block buffer
12
.
The two-dimensional DCT unit
13
reads the respective image signals from block buffer
12
by transforming the image signal corresponding to a plurality of image elements in respective blocks into the two-dimensional DCT coefficient corresponding to the space frequency of the same number as the plurality of image elements by using the two-dimensional discrete cosine transform.
Linear quantization unit
14
quantizes the two-dimensional DCT coefficient (matrix) of respective blocks by using the quantizing threshold value (quantizing matrix) which is suitable for vision and obtained from visual experimentation. The linear quantization unit
14
reads the quantizing threshold value from a quantizing threshold unit.
Variable length coding unit
160
performs a variable length coding of the quantization coefficient obtained by quantizing unit
14
by using the huffman coding table
17
obtained statistically, and outputs from the output terminal
18
the coded data obtained from the variable length coding method.
In the variable length coding method, the quantizing coefficient (matrix) arranged in a two-dimensional manner is converted to a one-dimensional train of numerical values by using a scanning operation called zig-zag scanning and performs a variable length coding based on the difference between the head DCT coefficient of a block and the head DCT coefficient of the preceding block with regard to the DC component. The value of the valid coefficients whose value is not zero is combined with the length of run (Run-length of the invalid coefficient whose value is zero), thereby applying the variable length coding to respective blocks.
FIG. 2
shows a structure of a conventional restoring apparatus of the static image for restoring data coded by the above-recited variable length coding to the images.
The data coded by the coding apparatus for the static image is inputted from the input terminal
21
to variable length restoring unit
22
in units of blocks.
The variable length restoring unit
22
restores the quantizing coefficient of the DC component and AC components in the respective blocks using the huffman coding table
17
(huffman decoding table
27
) which is used in the above variable length coding.
Dequantization unit
23
converts the decoded quantization coefficient of respective DC component and AC component into the two-dimensional DCT coefficients by using the quantized value
29
(quantized matrix).
Inverse DCT transforming unit
24
performs the two-dimensional DCT transform by using the two-dimensional DC coefficient of the respective blocks which are decoded and restored as the image elements.
The image signal of respective restored image elements is stored within the image memory
25
in units of blocks.
An operation of the image signal of respective image elements at image blocks into the image memory
25
is controlled by inside-block image element writing controlling unit
26
and controls the block address generating unit
28
and enables it to output an address of a block in which the image signal is to be written into the image memory
25
and performs a writing operation of the image signal of respective image elements of the present frame into the image memory
25
in units of blocks.
There is a method of coding a moving image which is subject to a continuous scene, i.e. a television conference or a television telephone system. Conventionally, the respective images are subject to the methods of coding the moving images.
FIG. 3
shows a block diagram of a basic part of the moving image coding apparatus by using an inter-frame prediction coding method of utilizing a relation between the image in reference frame and the image in the present frame.
The image of respective scenes of the moving image is inputted to difference image generating unit
320
through input terminal unit
31
.
The reference image which is obtained by restoring the image in the preceding frame and is used as the image in the reference frame.
The difference generating unit
32
obtains the difference image between the reference images stored in an image memory
33
and the present image (the image in the present frame) inputted from the input terminal
31
and stores the difference image in block buffer
34
. The difference image can be obtained for respective blocks comprising image elements of 8×8, for example. Therefore, the block buffer
34
stores the difference image i.e difference block image between the block in the reference frame of 8×8 image element, for example and the corresponding block in the present frame of 8×8 image elements with regard to the blocks in the same position.
The two-dimensional DCT transform
35
applies the two-dimensional DCT transform to the difference block image of 8×8 image elements stored in the block buffer
34
obtains the 8×8 DCT coefficients, and thereby outputs them to quantization unit
36
.
Quantization unit
36
applies a linear quantization to 8×8 coefficients by using adequate threshold value applicable to an eyesight, thereby providing 8×8 quantized coefficients.
The 8×8 quantization coefficients are subject to zig-zag scanning and is outputted to variable length coding unit
37
in the order of the DC component, AC components with a low space frequency AC components with relatively high space frequency.
The variable coding unit
37
performs a variable length coding of the 8×8 input quantized coefficients by using huffman coding table
38
and outputs the data coded by the variable length coding from the output terminal
40
.
The 8×8 quantizes coefficient matrix outputted from quantization
Murashita Kimitaka
Noda Tsugio
Fujitsu Limited
Lee Richard
Staas & Halsey , LLP
LandOfFree
Image data coding and restoring method and apparatus for... 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 data coding and restoring method and apparatus for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image data coding and restoring method and apparatus for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2579877