Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-01-28
2001-10-09
Le, Vu (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240120, C375S240130
Reexamination Certificate
active
06301301
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally relates to a coding technique for a moving image, and more particularly to a predictive coding technique for coding images by predicting a motion of an object.
2. Description of the Prior Art
FIG. 23
shows a structure of an moving image coding apparatus according to a first conventional technique, which is constructed based on the recommendation H.263 by ITU-T. In the figure, a reference numeral
1
is an input image,
101
is a differentiator,
102
is a prediction signal,
103
is a prediction error signal,
104
is an encoder,
105
is coded data,
106
is a decoder,
107
is a decoded prediction error signal,
108
is an adder,
109
is a local decoded image signal,
110
is a memory,
111
is a predictor, and
112
is a motion vector.
The input image
1
to be coded is input to the differentiator
101
. The differentiator
101
determines a difference between the input image
1
and the prediction signal
102
which will be described later, and outputs the difference as the prediction error signal
103
. The encoder
104
codes the input image
1
or the prediction error signal
103
and outputs the coded data
105
. The encoder
104
uses DCT (Discrete Cosine Transformation), which is one type of quadrature conversions, to convert the prediction error signal
103
from a space domain to a frequency domain, and linearly quantize conversion coefficients obtained by DCT
The coded data
105
obtained by the encoder
104
is then divided into two parts. One part is transmitted to a decoder on the receiver side (not shown), while the other is input to the decoder
106
provided in this apparatus. The decoder
106
operates in a reverse way to the encoder
104
, and determines and outputs the decoded prediction error signal
107
from the coded data. The adder
108
adds the decoded prediction error signal
107
to the prediction signal
102
, to thereby determine and output the decoded image signal
109
. The predictor
111
carries out motion compensative prediction using the input image
1
to be coded and the decoded image signal
109
of the frame located immediately previous to the current frame stored in the memory
110
, and outputs the prediction signal
102
and the motion vector
112
. Motion compensation is carried out by blocks of a fixed size. Each block is called a macro block consisting of 16×16 pixels. For those blocks located at heavy movement part of the image, an optional function is provided to carry out the motion compensative prediction by sub-blocks, each consisting of 8×8 pixels, formed by quartering one macro block. The thus obtained motion vector
112
is transmitted to a decoder on the receiver side (not shown), while the prediction signal
102
is sent to the differentiator
101
and the adder
108
.
FIG. 24
shows a structure of a moving image coding apparatus according to a second conventional technique. The apparatus is constructed based on the coding system proposed in “A Very Low Bit Rate Video Coder Based on Vector Quantization”, IEEE Trans. on Image Processing, Vol. 5, No. 2, February, 1996. In the figure, a reference numeral
113
is an area divider,
114
is a predictor,
115
is an area determinator,
116
is coding mode information,
117
is a motion vector,
118
is an encoder, and
119
is coded data.
As shown in
FIG. 24
, the system divides the input image
1
into a plurality of areas by the area divider
113
. The area divider
113
determines the area shape in accordance with a motion compensative prediction error. Specifically, among predetermined 10 block sizes, i.e., 4×4, 4×8, 8×4, 8×8, 8×16, 16×8, 16×16, 16×32, 32×16, and 32×32, the area divider
113
assigns smaller blocks to the areas including heavy movement, while assigning larger blocks to the areas including light movement, such as a background, by judging a distribution of inter-frame signals using a threshold value. More specifically, the area determinator
115
calculates a variance of the prediction error signal obtained by the predictor
114
and determines each block size based on the variance. Attribute information
116
, such as the area shape or the coding mode of each area, and the motion vector
117
are also determined at this stage. In accordance with the coding mode information included in the attribute information
116
, the encoder
118
codes the prediction error signal or the original signal to obtain the coded data
119
. The succeeding procedures are the same as the first conventional technique described above.
The first conventional technique restricts the area shape used as a coding unit to only two types which are both square. As a result, coding is carried out regardless of whether the area is located at little movement part, such as a background, or at heavy movement part, such as edge parts of an object. Thus, this technique is not capable of accomplishing the adaptive coding which dynamically corresponds to the scene configuration.
The second conventional technique provides square blocks of multiple sizes, so that low significance areas are coded coarsely by larger blocks and the high significance areas are coded finely by smaller blocks. Thus, coding can be carried out in accordance with the scene configuration of the image. However, as the area shape is limited to square blocks, this technique has to be improved to realize the adaptive coding corresponding to a desired area shape.
In addition, if coding is carried out for individual areas which are determined by the area division corresponding to a desired area shape, both the amount of calculations involved in the area division and the amount of codes required for the area shape will be increased.
SUMMARY OF THE INVENTION
The present invention is made to solve the above-mentioned problems. An object of the present invention is to provide a coding technique for a moving image by using an area dividing technique capable of precisely corresponding to various image configurations. More specifically, the present invention aims to attain an adaptive coding which better corresponds to the scene configuration of the image, while reducing the amount of calculations related to the area division as well as the amount of codes related to the coding mode, by selecting a suitable area dividing method or coding method in response to the contents of an image or individual areas. Another object of the present invention is to provide a decoding technique capable of precisely decoding the coded data of areas which have been divided into various shapes.
In one aspect of the present invention, a moving image coding apparatus which divides an input digital image into a plurality of areas with a predetermined method comprises: a memory which stores information representing the shape of the areas; a projector which reads from the memory the area shape of a previously coded image, and projects the area shape onto an image to be coded to obtain an initial area shape; a divider which further divides each area of the input image into smaller areas in accordance with a predetermined judging method, the initial area shape of the input image being determined by the projector; an integrator which integrates the areas divided by the divider while determining whether or not each area is integrated with an adjacent area in accordance with the predetermined judging method; and an encoder which codes information from each area, the information including the image signals, motion information, shape information, and other attribute information.
The divider of the moving image coding apparatus includes: an activity calculator which calculates an activity for each area determined by the projector; a comparator which compares the activity with a predetermined reference value; and a subdivider which divides a particular area having the activity greater than the reference value into smaller areas in accordance with the comparison result obtained by the comparator.
Also, the divi
Asai Kohtaro
Isu Yoshimi
Sekiguchi Shun-ichi
Le Vu
Mitsubishi Denki & Kabushiki Kaisha
LandOfFree
Apparatus and method for coding and decoding a moving 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 Apparatus and method for coding and decoding a moving image, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for coding and decoding a moving image will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2578530