Motion video signal processing for recording or reproducing – Local trick play processing – With randomly accessible medium
Reexamination Certificate
1999-11-29
2002-11-12
Boccio, Vincent (Department: 2615)
Motion video signal processing for recording or reproducing
Local trick play processing
With randomly accessible medium
C375S240010, C375S240210, C348S699000
Reexamination Certificate
active
06480670
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a video signal encoding method and system, and in particular to a video signal encoding method and system with motion compensated prediction.
A high-efficiency encoding system for use in encoding video signals employs a hybrid encoding system combining inter-picture prediction encoding utilizing motion compensation and intra-picture encoding.
FIG. 1
is a block diagram showing an encoding system utilizing a conventional hybrid encoding method described in ISO-IEC/JTC/SC29/WG11 MPEG 92/N0245 Test Model 2. As illustrated, a digital video signal
101
received at an input terminal
1
is supplied to a first input of a subtractor
10
, a first input of a motion compensated prediction circuit
17
, and a second input of a quantizer
12
. The output of the subtractor
10
is supplied to a DCT (discrete cosine transform) circuit
11
, and its output is supplied to a first input of the quantizer
12
. The output
102
of the quantizer
12
is supplied to a first input of a variable-length encoder
19
, and to an inverse quantizer
13
, and its output is supplied to an IDCT (inverse discrete cosine transform) circuit
14
, and its output is supplied to a first input of an adder
15
. The output of the adder
15
is supplied to a memory
16
, and data (reference image signal
103
) read from the memory
16
is supplied to a second input of the motion compensated prediction circuit
17
and a first input of a selector
18
. A first output
104
of the motion compensated prediction circuit
17
is supplied to the memory
16
.
A zero signal (data representing a value “0”) is supplied to a second input of the selector
18
, and a second output
107
of the motion compensated prediction circuit
17
is supplied to a third input of the selector
18
. The output
106
of the selector
18
is supplied to a second input of the subtractor
10
and a second input of the adder
15
. A third output
107
of the motion compensated prediction circuit
17
is supplied to a second input of the variable-length encoder
19
. The output of the variable-length encoder
19
is input to a transmitting buffer
20
, and a first output of the transmitting buffer
20
is output via an output terminal
2
. A second output
108
of the transmitting buffer
20
is supplied to a third input of the quantizer
12
.
FIG. 2
is a block diagram showing an example of configuration of a conventional motion compensated prediction circuit
17
. The digital video signal
101
is supplied to a first input of a motion vector search circuit
3
a
. A reference image signal
103
input from the memory
16
is supplied to a second input of the motion vector search circuit
3
a
. The motion vector
109
output from the motion vector search circuit
3
a
is supplied to a first input of a selector
4
a
. A zero vector (“0”) is supplied to a second input of the selector
4
a.
The prediction image
110
output from the motion vector search circuit
3
a
is supplied to a first input of a distortion calculator
5
a
. Applied to a second input of the distortion calculator
5
a
is the video signal
101
from the input terminal
1
. A distortion output
111
from the distortion calculator
5
a
is supplied to a first input of a comparing and selecting circuit
7
a.
The video signal
101
is also supplied to a first input of a distortion calculator
5
b
. The reference image signal
103
is also supplied to a second input of the distortion calculator
5
b
. A distortion output
112
from the distortion calculator
5
b
is supplied to a second input of the comparing and selecting circuit
7
a
. A selection mode
113
output from the comparing and selecting circuit
7
a
is supplied to a first input of a comparing and selecting circuit
7
b
. A distortion output
114
from the comparing and selecting circuit
7
a
is supplied to a second input of the comparing and selecting circuit
7
b.
The selection mode output
113
from the comparing and selecting circuit
7
a
is also supplied to a third input of the selector
4
a
. A motion vector
107
output from the selector
4
a
is supplied to the variable-length encoder
19
.
The prediction image
110
output from the motion vector search circuit
3
a
is supplied to a first input of a selector
4
b
. The reference image signal
103
from the input terminal
1
a
is also supplied to a second input of the selector
4
b
. The selection mode
113
from the comparing and selecting circuit
7
a
is supplied to a third input of the selector
4
b.
The prediction image
104
from the selector
4
b
is supplied to the memory
16
. The video signal
101
from the input terminal
1
is also input to a variance calculator
9
. An output
115
of the variance calculator
9
is supplied to a third input of the comparing and selecting circuit
7
b
. The selection mode
105
from the comparing and selecting circuit
7
b
is supplied to the selector
18
.
The operation is described next. The digital input signal
101
is supplied to the subtractor
10
, where a difference between the input picture (frame or field) and the picture from the motion compensated prediction circuit
17
is taken to reduce the temporal redundancy (redundancy in the direction of the time axis), and DCT is performed in the directions of the spatial axes. Coefficient obtained are quantized, and variable-length encoded, and then transmitted via the transmitting buffer
20
.
Motion compensated prediction is schematically illustrated in FIG.
3
. The picture that is to be encoded is divided into matching blocks each consisting of 16 pixels by 16 lines. For each matching block, examination is made as to which part of the reference picture, if used as a prediction image, minimizes the distortion. For instance, in the case of a still picture, if the 16 pixels by 16 lines at the same position as the matching block are used as the prediction image, the distortion will be zero. In the case of a motion picture, it may be that the block shifted leftward by 8 pixels and downward by 17 lines for instance yields the minimum distortion. Then, this block at the shifted position is regarded as a block corresponding to the matching block in question, and used as the prediction image, and (−8, 17) is transmitted as the motion vector.
Further explanation of the motion compensated prediction is explained with reference to FIG.
2
. First, in the motion vector search circuit
3
a
, the motion vector is determined on the basis of the input image
101
and the reference image
103
. This is effected by finding a block in the reference picture which minimizes the distortion for each matching block, as explained in connection with
FIG. 3
, and the the block thus found to give the minimum distortion is used as the prediction image, and the position of the block thus found to give the minimum distortion relative to the matching block is used as the motion vector. The distortion may be defined in terms of the sum of the absolute values of the differences.
In the distortion calculator
5
a
, the distortion defined as the sum of the squares of the differences between the input image
101
and the prediction image
110
output from the motion vector search circuit
3
a
is calculated for each matching block. The distortion
111
is also denoted by SEmc. In the distortion calculator
5
b
, the distortion defined as the sum of the squares of the differences between the input image
101
and the reference image
103
(of the same position) is calculated for each matching block. This distortion
112
is also denoted by SEnomc. The SEnomc is a particular value of the distortion SEmc where the vector representing the relative position between the input image
101
and the prediction image is zero.
For the purpose of the following explanation, it is assumed that the whole picture consists of I pixels by J lines, and the input picture is represented by F(i,j) where i represents the pixel number in the horizontal direction and 0≦i<I, and j represents the pixel number in the vertical direction and 0≦j&l
Hasegawa Hiromu
Hatano Yoshiko
Kasezawa Tadashi
Mishima Hidetoshi
Okazaki Koji
Birch & Stewart Kolasch & Birch, LLP
Boccio Vincent
Mitsubishi Denki & Kabushiki Kaisha
LandOfFree
Video signal encoding method and system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Video signal encoding method and system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Video signal encoding method and system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2934577