Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-11-09
2004-09-28
Lee, Young (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
Reexamination Certificate
active
06798834
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an image coding apparatus which is used in image communications equipment such as, visual telephone and teleconferencing equipment and image storage/recording equipment such as a digital VTR.
The invention also relates to a motion prediction circuit which performs motion detection and motion prediction on image information as well as to image coding and decoding apparatuses to which the above motion prediction circuit is applied.
2. Description of the Related Art
Conventionally, various image coding techniques have been developed to improve the efficiency of transmission and recording of digital images.
FIG. 9
is a block diagram showing a prior art image coding apparatus that is shown in Recommendation H.261 of ITU-T (International Telegraph and Telephone Consultative Committee). In this apparatus, first an input digital image signal
101
to be coded is input to a differentiator
21
. The differentiator
21
takes a difference between the input digital image signal
101
and a prediction signal
136
(described later), and outputs a resulting difference signal as a prediction error signal
131
. A coding section
22
encodes the prediction error signal
131
and outputs resulting code data
132
. The above Recommendation employs, as a coding method in the coding section
22
, a technique of converting the prediction error signal
131
from the spatial domain to the frequency domain by using DCT (discrete cosine transform) that is a kind of quadrature transform, and linearly quantizing resulting transform coefficients.
The code data
132
as output from the coding section
22
in branched into two parts, one of which is sent to a reception-side decoding apparatus (not shown). The other part is input to a decoding section
23
of the image coding apparatus under discussion. The decoding section
23
performs an inverse operation to the operation of the coding section
22
, that is, it produces a decoded prediction error signal
133
based on the code data
132
. An adder
24
adds the decoded prediction error signal
133
to the prediction signal
136
, to thereby produce a decoded image signal
134
. One frame of the image signal
134
is stored in a memory
25
such as a frame memory. The memory
25
outputs the stored decoded image signal of one frame after delaying it by one frame, as a decoded image signal
135
of a preceding frame. A predictor
26
produces the prediction signal
136
and a motion vector
137
by performing motion-compensating prediction based on the input digital signal
101
to be coded and the 1-frame preceding decoded image signal
135
. The motion vector
137
is sent to the reception-side decoding apparatus (not shown) and the prediction signal
136
is supplied to the differentiator
21
and the adder
24
.
In the conventional image coding apparatus having the above configuration, the coding section
22
encodes an image of one frame substantially uniformly irrespective of the content of a subject image. Further, the image
25
operates such that an image of only one frame is stored therein and rewritten every frame. Therefore, the efficiency of coding cannot be improved in the conventional image coding apparatus.
A prior art motion prediction circuit is disclosed in Japanese Unexamined Patent Publication No. Hei. 4-347987, which is shown in FIG.
19
. In
FIG. 19
, reference numeral
611
denotes a motion vector detecting circuit;
612
, a segmenting circuit; and
613
, a motion parameter detecting circuit. Further, reference numeral
601
denotes an input image signal;
602
, interframe motion vectors;
603
, segmentation information; and
604
, a motion parameter that is detected on a-segment-by-segment basis.
The above circuit operates in the following manner. An input image signal
601
is supplied to both the motion vector detecting circuit
611
and the segmenting circuit
612
. The motion vector detecting circuit
611
detects a motion vector
602
on a pixel or small-block basis and supplies the detected motion vectors
602
to the motion parameter detecting circuit
613
. On the other hand, the segmenting circuit
612
divides the input image signal
601
into a plurality of segments having different motion by dividing or combining blocks by referring to the motion vectors
602
, and outputs segmentation information
603
indicating a manner of segmentation. The motion parameter detecting circuit
613
selects motion vectors belonging to a segment from the motion vectors
102
, and calculates first-order conversion coefficients which describe motion of the segment based on the selected motion vectors. By performing this operation for all the segments, the motion parameter detecting circuit
613
outputs detected motion parameters
604
as motion information of the input image.
FIG. 20
shows details of the motion parameter detecting circuit
613
. First, coordinates of N measurement points belonging to a target segment of motion parameter detection are selected. The N sets of measured coordinates are supplied to a center-of-gravity detecting circuit
621
, which calculates coordinates of the center of gravity of the N measurement points.
The center-of-gravity detecting circuit
621
then converts the coordinates of the N measurement points into coordinates relative to the calculated center of gravity, and supplies the resulting coordinates to average detecting circuits
622
-
624
. Further, x-components Vx(X, Y) of the N measured motion vectors are input to the average detecting circuit
622
, and y-components Vy(X, Y) are input to the average detecting circuit
624
. The average detecting circuits
622
-
624
calculate various averages and supply those averages to first-order conversion coefficients detecting circuits
625
and
626
. The first-order conversion coefficients detecting circuit
625
calculates first-order conversion coefficients a, b, and e by a method of least squares based on the received averages. Similarly, the first-order conversion coefficients detecting circuit
626
calculates first-order conversion coefficients c, d, and f. These coefficients are used as motion parameters of the segment concerned. Thus, the motion parameters of each segment of the input image are obtained.
Another conventional motion prediction circuit is disclosed in Japanese Unexamined Patent Publication No. Hei. 5-328334. This publication discloses a block coding scheme in which coding is performed in units of enlarged blocks each obtained by combining similar adjacent blocks.
In the first conventional example described above, a method of least squares is used in converting the motion vectors
602
into the motion parameters
604
, i.e., first-order conversion coefficients. This is based on the assumption that the motion of a segment is sufficiently small. Therefore, detection errors become large when the motion is large.
Further, coordinates of N measured points belonging to the target segment are needed to produce the motion parameters
604
. In the case of real-time communication of moving images, this requires that motion vectors of each segment of an input image be measured automatically at high speed with high accuracy, which is not realistic.
Still further, there is no criterion to judge whether optimum segments have been obtained for the motion vectors detecting operation. Therefore, the prediction performance strongly depends on the accuracy of segmentation.
On the other hand, in the second conventional example described above, coding is performed on a block-by-block basis in which a plurality of blocks are combined into a larger block in a fixed manner. Therefore, the prediction error of an image cannot be made smaller than a block.
SUMMARY OF THE INVENTION
The present invention has been made to solve the above problems in the art and, therefore, has an object of providing an image coding apparatus which can improve the efficiency of coding by coding an image in accordance with its content and rewriting the contents of memories in accordance with
Asai Kohtaro
Fukuhara Takahiro
Murakami Tokumichi
Yamada Yoshihisa
Lee Young
Mitsubishi Denki & Kabushiki Kaisha
LandOfFree
Image coding apparatus with segment classification 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 Image coding apparatus with segment classification and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image coding apparatus with segment classification and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3242649