Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-03-25
2001-04-03
Kelley, Chris (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240120
Reexamination Certificate
active
06212236
ABSTRACT:
BACKGROUND OF THE INVENTION
(1) Field of the Invention
This invention relates to an image decoding apparatus for decoding coded data to generate moving pictures, the data having been coded in conformity with a standard such as Moving Picture Experts Group (MPEG).
(2) Description of the Prior Art
Recently, great efforts have been made to develop efficient techniques for coding/compressing moving pictures. These techniques, such as MPEG, are used in the fields of computers, communications, broadcasting, etc.
According to MPEG, the moving pictures are divided into frames, each frame being divided into a plurality of blocks, and each block generally including 64 pixels (8 pixels×8 pixels). The moving picture data is encoded block by block for each frame. The encoding procedure includes processes such as motion estimation, Discrete Cosine Transformed (DCT) being an orthogonal transformation, quantization, and entropy coding. The original moving picture data changes into coded data (bitstreams) through the above processes. Moving pictures are restored from the bitstreams by going through a reversed procedure of the coding procedure which includes entropy decoding (dividing and decoding of the bitstreams), dequantization, inverse DCT, and motion compensation.
The following is a description of the principle of “predictive coding” cited from “Gazou Deta Asshuku No Kiso Chishiki (Basic Knowledge of Image Data Compression),” Interface, December 1991, CQ Publishing Co. Predictive coding is a method of predicting current pixel values from preceding pixel values using a difference between a true value and a predictive value, the difference being called a predictive error. The method uses a unique characteristic of image data that the luminances of adjacent pixels strongly correlate.
The
FIG. 1
, pixel (luminance) value sequence
1112
(0, 0, 0, 1, 0, 3, 1, 0, 1, 1, . . . ) are actual values of original pixel value Xi
1102
which are processed in sequence. The luminance appears to increase gradually from the left-hand side to the right-hand side in this case. In the present description, it is assumed that each pixel holds a pixel (luminance) value ranging from “0” to “15.” The numbers in the column of pixel number i
1101
represent serial numbers assigned to pixels processed in sequence. As shown in the drawing, the sequence
1112
of the original pixel value X
1
1102
does not appear to change greatly in each pair of adjacent pixels. By using this characteristic, it is possible for a receiver to approximately predict each of the next pixel values. It is further possible for a transmitter to assume that the receiver will predict each next pixel value. As a result, after sending an original pixel value X
1
1102
, the transmitter transmits a difference D
2
=X
2
Y
1
, where Y
1
represents a predictive value (−X
1
), and X
2
represents a true value. The receiver obtains the true value X
2
by adding the difference D
2
to the pixel value X
1
.
The transmitter regards a value X
i−1
as a predictive value Yi
1103
(Yi−X
i−1
). The transmitter calculates predictive difference Di
1104
between predictive value Yi
1103
and true value Xi
1102
. Thus, Di=Xi−Yi. The transmitter sends the calculated predictive difference Di
1104
as transmission value Ti
1105
to the receiver via transmission path
1106
. The receiver receives the transmission value Ti
1105
as reception value Ri
1107
(Ri=Ti). The receiver generates decoded value Zi
1108
by adding preceding decoded value Z
i−1
to the reception value Ri
1107
(Zi=Z
i−1
+Ri). After going through the above procedure, a decoded-value sequence
1118
(0, 0, 0, 1, 0, 3, 1, . . . ) is generated. Note that a pixel holding the pixel value X
i−1
preceding the original pixel value Xi is called a reference pixel.
The above method is the simplest one in which a preceding pixel is treated as a predictive value. This type of predictive coding method using pixel values of preceding pixels is called forward predictive coding, a predictive coding method using a succeeding pixel is called backward predictive coding; and a predictive coding method using a preceding pixel and a succeeding pixel is called bidirectional predictive coding.
The above predictive coding applied to adjacent pixels in a frame is called intra predictive coding. Also, this predictive coding applied to frames of moving picture data is called inter predictive coding.
In the above description of the principle, the predictive coding is performed for each pixel. In general, however, the predictive coding is performed for each block of 8 pixels×8 pixels. In this case, when a block is the same as the preceding block (called a reference block), Information indicating that these blocks are the same may be sent instead of 64 difference values “0”, reducing the amount of transmitted information.
FIG. 2
shows a hierarchical structure of coded data which is generated by coding (compressing) moving picture data with a moving picture compression technique. As shown in a sequence layer
1201
, a code sequence
1211
corresponding to a piece of moving picture data is divided into a plurality of Groups Of Pictures (GOP)
1212
. Other information such as Sequence Header Code (SHC) is attached to each GOP when the codes are transmitted. As shown in a GOP layer
1202
, each GOP
1212
is composed of a Group Start Code (GSC) being a start code of the GOP and a plurality of Intra-Coded Pictures (I pictures)
1221
, Bidirectionally Predictive-Coded Pictures (B-pictures)
1222
, and Predictive-Coded Pictures (P-pictures)
1223
.
Each of I-pictures
1221
, B-pictures
1222
, and P-pictures
1223
includes the same amount of data as one frame.
The I-picture is a result of coding only one frame without obtaining a difference between the frame and other frames. The P-picture is a result of a predictive coding and includes difference values obtained from a calculation using pixel values of the current frame and the preceding frame. The B-picture is also a result of a predictive coding and includes difference values obtained from a calculation using pixel values of the current frame and the preceding and succeeding frames. As a result of this, in the decoding process, the preceding and succeeding frames of a B-picture must first be decoded before the B-picture itself is decoded. Similarly, a preceding frame should be decoded before a target P-picture is decoded.
As shown in picture layer
1203
, each picture is composed of a Picture Start Code (PSC)
1233
for specifying a picture type of I-picture, P-picture, or B-picture, a Picture Coding Type (PCT)
1232
, and a plurality of slices
1231
.
Each slice
1231
corresponds to one of pixel sequences making up a horizontal line in a frame.
As shown in slice layer
1204
, each slice
1231
is composed of a Slice Start Code (SSC) for indicating the start of the slice layer, and a plurality of Macroblocks (MB)
1241
.
As shown in macroblock layer
1205
, each macroblock
1241
is composed of a plurality of Blocks (B)
1251
and information such as MacroBlock Type (MBT)
1615
specifying a macroblock type such as I-picture, P-picture, or B-picture, Motion Horizontal Forward Code (MHF)
1252
indicating a horizontal element of a forward motion vector of macroblock
1241
, Motion Vertical Forward Code (MVF)
1253
indicating a vertical element of the forward motion vector of macroblock
1241
, Motion Horizontal Backward Code (MHB)
1254
indicating a horizontal element of a backward motion vector of macroblock
1241
, Motion Vertical Backward Code (MVB)
1255
indicating a vertical element of the backward motion vector of macroblock
1241
, and Coded Block Pattern (CBP)
1256
specifying a pattern of six blocks included in microblock
1241
.
The blocks
1251
are generally composed of six blocks
1261
,
1262
,
1263
,
1264
,
1265
, and
1266
. Blocks
1261
,
1262
,
1263
, and
1264
are each composed of an element specifying a luminance, and blocks
1265
and
1266
are e
Hirai Makoto
Kimura Kozo
Kiyohara Tokuzo
Nishida Hideshi
Kelley Chris
Matsushita Electric - Industrial Co., Ltd.
Price and Gess
Wong Allen
LandOfFree
Image 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 Image decoding apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image decoding apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2527250