Image analysis – Image compression or coding – Including details of decompression
Reexamination Certificate
1998-09-01
2002-05-21
Tran, Phuoc (Department: 2621)
Image analysis
Image compression or coding
Including details of decompression
C382S236000
Reexamination Certificate
active
06393153
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an apparatus for decoding, through expansion, a moving picture signal compressed through two-dimensional picture compression, in particular, by a data compression technology using an interframe prediction as in the so-called MPEG system.
2. Description of Related Art
The data compression conforming to the so-called MPEG/Video (will be referred to as “MPEG” hereinafter) being the international standard for the moving picture compression, involves three types of frames (pictures) which will be described below. One of them is an I-picture (intra-code picture). This frame is not interframe predicted but compressed within itself. Therefore, the I-picture can be decoded requiring no data of any other frame. Other two frames are P-picture and B-picture. The P-picture is a predictive-coded picture, and B-picture is a bidirectionally predictive-coded picture. The P-picture is decoded through forward prediction from an I- or P-picture existing before the P-picture or frame in consideration. The B-picture is decoded through backward prediction from a frame existing after the frame in consideration in addition to forward prediction from an I- or P-picture existing before the frame in consideration. That is, the B-picture can be decoded through any of the forward, backward and bidirectional predictions. The MPEG referred to herein stands for the Moving Picture-coding Experts Group of the ISO/IEC JTC/SC29 (International Organization for Standardization/International Electrotechnical Commission, Joint Technical Committee 1/Sub Committee 29). The MPEG-1 (MPEG phase 1) standard includes ISO 11172, and the MPEG-2 standard includes ISO 13818. In these international standards, ISO 11172-1 and ISO 13818-1 are included in the item “system multiplexing”, ISO 11172-2 and ISO 13818-2 are included in the item “video coding”, and ISO 11172-3 and ISO 13818-3 are included in the item “audio coding”.
FIG. 1
shows an example of the dependency of the above three types of frames I, P and B upon each other.
In
FIG. 1
, frames (pictures) are indicated with references I
1
, B
2
, B
3
and P
4
, respectively, numbered from the left to right. As mentioned above, “I” means an intra-coded picture, namely, I-picture, “P” indicates a predictive-coded picture, namely, P-picture, and “B” indicates a bidirectionally predictive-coded picture, namely, B-picture. The arrows indicate directions of prediction, respectively. I
1
is an I-picture numbered one, and decoded within itself. P
4
is a P-picture numbered four. For decoding this frame, it is necessary to use a motion vector for reading the result of I
1
frame decoding for calculation. B
2
is a B-picture numbered
2
. It is decoded through bidirectional prediction. For image restoration, an interframe prediction (that is, motion compensation) is done based on the results of I
1
frame decoding and P
4
frame decoding.
FIG. 2
shows, along the time base, necessary data for decoding frames of the above types.
In
FIG. 2
, data of the frame I
1
is required for decoding itself as well as for motion compensation (MC) for each of the frames B
2
, B
3
and P
4
. For decoding B-pictures including the frames B
2
, B
3
, etc., data of the frame P
4
is required in addition to the data of the B-pictures themselves. The frame P
4
being a P-picture is necessary for decoding the frames B
2
, B
3
, P
4
, B
5
and B
6
as shown in FIG.
2
. The I-picture and P-picture required for decoding the B-pictures should have been decoded before the B-pictures are decoded, and should be held all the way through the decoding of the B-pictures. As seen from
FIG. 2
, data for up to three frames including the B-picture have to be held at a time. The buffer memories for holding these frames should have a large storage area. Therefore, they are used as external memories incidentally to the decoder.
FIG. 3
shows an example of the time of decoding in the actual MPEG decoder.
In
FIG. 3
, a time of decoding is indicated with a reference T
1
while a time of presentation is indicated with a reference T
2
. In this example, a time gap between the decoding and presentation equivalent to 1.5 frames for I-picture (as indicated with T
3
in FIG.
3
), and 0.5 frame for B-picture (as indicated with T
4
). As seen also from
FIG. 2
, the result of I
1
decoding in
FIG. 3
is not only required for decoding the I
1
frame, but for decoding the frames P
4
, B
2
and B
3
, and so should be held for the time of their decoding. The P
4
frame is necessary for presentation of the frame P
4
itself and decoding of frames B
2
, B
3
, P
7
, B
5
, etc., and so held for the time of their decoding. The B-pictures are held either by buffering one frame or by holding only a part thereof, any of which will require a large memory area.
FIG. 4
shows an example of the conventional MPEG decoder configuration.
In
FIG. 4
, an MPEG bit stream is supplied to an MPEG decoder
7
at a terminal
6
thereof. Then, it is first decomposed by a demultiplexer
8
into audio and video bit streams. The bit streams are passed over a signal line
9
, bus
10
and signal line
11
, and written once into an area
14
of an external memory
13
via a memory interface circuit
12
. There is shown a signal line
15
which represents a data bus, address signal line, etc. laid between the decoder
7
and external memory
13
. Note that a block for processing audio signals is not shown in FIG.
1
.
The video bit stream read from the area
14
of the external memory
13
is passed through the memory interface circuit
12
, signal line
11
, bus
10
and a signal line
16
to a video decoder
17
.
The video decoder
17
comprises mainly a header/VLD circuit
18
for analysis of header and variable-length code, inverse quantizer (IQ; also dequantizer)
19
, inverse discrete cosine transformer (IDCT)
20
, motion compensator (MC)
21
, etc.
The motion compensator
21
is supplied via a signal line
22
with a reference image read using the motion vector and reconstructs or restores it. The reconstructed or output image from the motion compensator
21
is passed through a signal line
23
, bus
10
, signal line
11
and memory interface circuit
12
, and written into an area in the external memory
13
. For example, a decoded I-picture, P-picture and B-picture are written into designated areas
24
,
25
and
26
, respectively, in the external memory
13
. Each of these areas should have a size for one frame. The external memory
13
should have a rather large size. Note that the area
24
for the I-picture is also available for the P-picture.
The data of the I-picture in the external memory
13
is read at the time of presentation in
FIG. 3
, passed to an image display circuit
28
over a signal line
27
, and delivered as an image data at a terminal
29
. Also, the data of the I-picture in the external memory
13
is also read for decoding the B- or P-picture at the time shown in FIG.
3
. The data of the P-picture is similarly read for presentation and also for decoding of the B-picture. In
FIG. 4
, the B-picture is shown written in one-frame area in this example. It is read from this area at the time of presentation shown in FIG.
3
.
The aforementioned MPEG system combines together a variety of data compressing techniques to implement a high quality and compressibility of the video compression. The interframe prediction is one of the most important techniques. In the interframe prediction system, a motion vector is used to extract a block most approximate to a block being currently decoded from a frame existing before or after the frame in consideration, calculate a difference between the blocks and compress the result of the calculation. To decode a bit stream compressed through the interframe prediction, however, a frame data having previously been decoded should be held, which will require an increased memory capacity.
More particularly, the above-mentioned example of the conventional MPEG decoder needs to store decoded image data for at least 3 frames
Alavi Amir
Frommer William S.
Frommer & Lawrence & Haug LLP
Kessler Gordon
Sony Corporation
LandOfFree
Compressed moving picture signal decoding apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Compressed moving picture signal decoding apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Compressed moving picture signal decoding apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2875945