Pulse or digital communications – Bandwidth reduction or expansion
Reexamination Certificate
1998-10-08
2002-09-17
Kelley, Chris (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
C382S232000, C382S233000
Reexamination Certificate
active
06452968
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to circuitry, an apparatus and a method for the conversion of a decoded image and also to a data recording medium capable of being read by a computer and storing a decoded image conversion program. The present invention is applicable to circuitry and an apparatus for decoding data coded by an MPEG (Moving Picture Coding Experts Group) 2 or similar standardized system and displaying an image represented by the decoded data.
2. Description of the Background Art
It is a common practice with an image coding system based on the MPEG2 system to divide one frame of picture into macroblocks each having 16×16 pixels. Redundant information included in the frame are compressed in the time domain macroblock by macroblock. In addition, to implement a random access function and a high coding efficiency, the MPEG2 type system defines three different picture types, i.e., an I (Intra coded) picture, a P (Predictive coded) picture, and a B (Bidirectionally predictive coded) picture on a frame basis.
An I picture is coded within a frame (intra coding) independently of the other pictures. All the macroblocks constituting an I picture are subjected to intra coding. By arranging such I pictures periodically, it is possible to effect random access or to use I pictures as error recovery pointers. While frequent appearance of I pictures lowers the total coding efficiency of the coding system, a higher coding efficiency is achievable with I pictures at scene switchover points and for images with a low prediction efficiency. A P picture is a forward predictive coded picture, i.e., subjected to predictive coding using an I picture or a P picture appeared in the past. A B picture is a bidirectionally predictive coded picture, i.e., subjected to predictive coding using I pictures or P pictures included in one or both of two pictures preceding and following the B picture; the direction of prediction is selected macroblock by macroblock. Even in a region where a certain object appears or disappears, B pictures allow predictive coding to be accurately effected by use of preceding and/or following pictures and noticeably enhance the coding efficiency.
I, P and B pictures may be combined in any desired format feasible for the object of a coding apparatus. As for the coding order, I pictures and P pictures following B pictures with respect to time are coded before the B pictures.
FIG. 2
shows a specific original image represented by a sequence of pictures B
0
, B
1
,
12
, B
3
, B
4
, P
5
, B
6
, B
7
, P
8
, B
9
, B
10
, P
11
, B
12
, B
13
and B
14
. These pictures are coded in the order of
12
, B
0
, B
1
, P
5
, B
3
, B
4
, P
8
, B
6
, B
7
, P
11
, B
9
, B
10
, P
14
, B
12
and B
13
, as shown below the above original sequence in FIG.
2
. The coded picture sequence is written to a recording medium and then decoded in the same order as it has been coded, as shown in the middle part of FIG.
2
. Finally, the decoded sequence is rearranged in the original order, i.e., B
0
, B
1
,
12
, B
3
, B
4
, P
5
, B
6
, B
7
, P
8
, B
9
, B
10
, P
11
, B
12
, B
13
and B
14
and then displayed as a reproduced image.
FIG. 3
shows conventional decoding circuitry based on the MPEG2 system. As shown, a coded bit stream coming in through an input port
40
is written to a receipt buffer
41
. A VLC (Variable Length Code) decoder
42
separates the coded data into various kinds of data. Among the separated data, quantized DCT (Discrete Cosine Transform) coefficients are dequantized by a dequantizer
43
to turn out DCT coefficients. The DCT coefficients are subjected to inverse DCT transform by an inverse DCT
44
. If the image data are representative of an I picture, then they are directly written to one of two frame memories (FM
1
and FM
2
)
45
and
46
.
Assume that the image data output from the inverse DCT
44
are representative of a P picture. Then, image data subjected to motion compensation by a motion compensation circuit (MC
1
)
47
in accordance with motion vectors are read out of the frame memory
45
. Of course, such data may be those subjected to motion compensation by the other motion compensation circuit (MC
2
)
48
and read out of the other frame memory
46
. The image data read out of the frame memory
45
are selected by a switch
34
and added, pixel by pixel, to the image data output from the inverse DCT
44
by an adder
50
. The image data output from the adder
50
are written to the other frame memory
46
. When the image data output from the inverse DCT
44
are representative of a B picture, image data respectively subjected to motion compensation by the motion compensation circuits
47
and
48
in accordance with motion vectors are respectively read out of the frame memories
45
and
46
. In this. case, an averaging circuit (A)
49
averages the image data read out of the frame memories
45
and
46
and outputs the resulting mean values in the form of pixel values. The pixel values are selected by the switch
34
and added, pixel by pixel, to the image data output from the inverse DCT
44
by the adder
50
. However, the reproduced image derived from the B picture is written neither to the fame memory
45
nor to the frame memory
46
. The frame memories
45
and
46
are assumed to be used alternately.
The B picture reproduced by the above processing is directly output via an output port
36
. The I and P pictures also reproduced by the above processing each is read out of the associated frame memory
45
or
46
and then output via the output port
36
. As a result, a reproduced image is successfully displayed in the same order as the original image. A switch
35
selects one of the outputs of the adder
50
, and frame memories
45
and
46
at a time.
In practice, however, the decoding circuitry based on the MPEG2 system is usually constructed to output display image data in the interlacing order. Therefore, the decoding circuit of the type outputting image data in the form of macroblocks must be followed by a memory for macroblock/raster scan conversion. Moreover, if the picture coding type has a frame structure, as distinguished from a field structure, then the above memory must be capable of storing at least half a field of image data when image data are output by interlacing.
To solve the above problem, Japanese patent laid-open publication No. 59084/1995 discloses an image processing system capable of dealing with compressed image data fed thereto in the form of packets. Specifically, as shown in
FIG. 4
, the image processing system taught in the above document includes three frame memory areas FM
1
, FM
2
and FM
3
exclusively assigned to I pictures, P pictures and B pictures, respectively. With these frame memory areas FM
1
, FM
2
and FM
3
, it is possible to output image data in the interlacing order even when the picture coding type has a frame structure. More specifically, the image processing system uses four picture pointers RP, FP, BP and DP (see
FIGS. 5 and 6
) in order to determine the frame memory area FM
1
, FM
2
or FM
3
which a memory controller MCU should access. The picture pointers RP, FP, BP and DP respectively point at the frame memory areas where a current reproduced picture, a forward picture, a backward picture, and a current display picture are stored.
FIG. 5
demonstrates the transition of the picture pointers RP, FP, BP and DP occurring when pictures of different types are decoded and displayed. At the time when the first picture I
0
is decoded, the image is not displayed yet. At this instant, the reproduced picture pointer RP points at an unoccupied area for storing the picture I
0
, e.g., the frame memory area FM
1
.
When a picture P
1
following the picture I
0
is decoded, the picture I
0
is displayed without fail. At this instant, the reproduced picture pointer RP points at, e.g., the frame memory area FM
2
while the display picture pointer DP points at the frame memory area FM
1
storing the picture I
0
. Because a predictive picture for P
1
need the for
An Shawn S.
Frank Robert J.
Kelley Chris
Oki Electric Industry Co. Ltd.
Sartori Michael A.
LandOfFree
Circuitry and apparatus for converting a decoded image with... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Circuitry and apparatus for converting a decoded image with..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Circuitry and apparatus for converting a decoded image with... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2855138