Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-06-18
2001-08-14
Kelley, Chris (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
Reexamination Certificate
active
06275533
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to an image processing method, an image processing apparatus, and a data recording medium and, more particularly to an image processing method, an image processing apparatus, and a data recording medium in which adaptive intra-frame prediction is performed to frequency components of an interlaced image signal, thereby improving efficiency in coding of an image signal.
BACKGROUND OF THE INVENTION
Predictive coding which compresses image data of a moving picture using its redundancy, includes intra-frame predictive coding which predicts image data using image data in a target frame to-be-coded (hereinafter referred to as a target frame), and inter-frame predictive coding which predicts image data using image data in a frame other than the target frame.
More specifically, in the intra-frame predictive coding, a prediction value of image data in a target frame is generated from the image data in the target frame, and a difference value between a value of the image data and the prediction value is coded, thereby eliminating or reducing a large amount of spatially redundant information inherent in image data, when compressing image data.
Meanwhile, in the inter-frame predictive coding, a prediction value of image data in a target frame is generated from data in another frame, and a difference value between a value of the image data and the prediction value is coded, thereby eliminating or reducing a large amount of temporarily redundant information included in image data of an image in small motion, when compressing image data.
Recently, DCT (Discrete Cosine Transform) has been widely used in image coding. According to MPEG (Moving Picture Expert Group) as a typical image coding method, an image space (frame) formed by a digital image signal is divided into plural rectangular regions (blocks) as DCT units, and the DCT is performed to the image signal for each block.
An intra-frame prediction method of DCT coefficients, i.e., image data in a DCT domain (frequency domain) which is adopted by MPEG is described in “Intra DC and AC prediction for I-VOP and P-VOP” of ISO/IEC JTC1/SC29/WG11 MPEG97/N1642 MPEG-4 Video Verification Model Version 7.0 (hereinafter referred to as MPEG-4 VM 7.0).
According to this reference, a DC (direct current) component and AC (alternating current) components of DCT coefficients of a target block to-be-coded (hereinafter referred to as a target block) are predicted using DCT coefficients of three blocks positioned upper left, above, and left, with respect to and adjacently to the target block in an image space.
FIG. 15
is a diagram for explaining an intra-frame DCT coefficient prediction method described in the reference which is adopted by a prior art image coding method.
Referring now to
FIG. 15
, there are shown 4 blocks (DCT blocks) R
0
-R
2
, and X as DCT units, each comprising 8×8 pixels. These blocks are positioned adjacently to each other in an image space (spatial domain) formed by an image signal. Assume that the block X is a target block, and the blocks R
0
, R
1
, and R
2
are blocks which have been coded (hereinafter referred to as a coded block), which are positioned upper left, above, and left, with respect to and adjacently to the target block in the spatial domain.
In the prior art intra-frame DCT coefficient prediction method, prediction values of DCT coefficients of the block X are generated with reference to DCT coefficients of the block R
1
or the block R
2
.
Specifically, where DCT coefficients of the coded block R
1
are referred to, a DC component in the upper left corner and AC components in the highest row of the coded block R
1
are used as prediction values of DCT coefficients positioned at the corresponding positions of the block X. Where DCT coefficients of the coded block R
2
are referred to, a DC component in the upper left corner and AC components in the leftmost column of the coded block R
2
are used as prediction values of DCT coefficients positioned at the corresponding positions of the block X.
Decision on whether to refer to the coded block R
1
or R
2
, is made using DC components of the coded blocks R
0
, R
1
, and R
2
.
Where an absolute value of difference of DC components between the blocks R
0
and R
2
is smaller than an absolute value of difference of DC components between the blocks R
0
and R
1
, there is high correlation of DCT coefficients between blocks arranged in the vertical direction, and therefore the DCT coefficients of the block R
1
are referred to. On the other hand, where the absolute value of difference of the DC components between the blocks R
0
and R
1
is smaller than the absolute value of difference of the DC components between the blocks R
0
and R
2
, there is high correlation of DCT coefficients between blocks arranged in the horizontal direction, and therefore the DCT coefficients of the block R
2
are referred to.
As described in “Adaptive Frame/Field DCT” of the reference (MPEG-4 VM7.0), DCT (frequency transformation) for use by coding process of an interlaced image, comprises two types of DCT, namely, frame DCT and field DCT. In the frame DCT, image data is transformed frame by frame, while in the field DCT, image data is transformed field by field. According to MPEG, switching between the frame DCT and the field DCT is adaptively performed for each macroblock comprising 4 blocks.
This switching is performed depending on whether or not scanning lines have been rearranged as shown in FIG.
16
. In the field DCT, DCT is performed to image data of each of 4 blocks of a macroblock in which scanning lines have been rearranged.
More specifically, in the frame DCT, DCT is performed to image data of each of 4 blocks of macroblock in which even-numbered scanning lines and odd-numbered scanning lines are alternately arranged. In the field DCT, scanning lines are rearranged, and thereby a macroblock comprises a first field block comprising even-numbered scanning lines, and a second field block comprising odd-numbered scanning lines, followed by performing DCT to image data of each block of the macroblock.
Thus, in coding process of an interlaced image signal, macroblocks to which the frame DCT is performed and macroblocks to which the field DCT is performed coexist in an image space.
In a case where correlation of pixel values between the first and second fields is higher than those in the first and second fields, the frame DCT is performed, and in other cases, the field DCT is performed.
Accordingly, in some cases, DCT of neighboring macroblocks or adjacent blocks (subblocks of a macroblock) differ from each other. In such cases, with respect to DCT coefficients, there is correlation between adjacent macroblocks or blocks lower than in the case of using one type DCT.
In other cases, fields to which adjacent blocks belong, differ from each other between these blocks. Also in such cases, with respect to DCT coefficients, there is correlation between adjacent blocks lower than in a case where the blocks belong to the same field.
However, since in macroblocks which have been processed by the field DCT (field DCT type macroblocks), first field blocks and second field blocks coexist, it is difficult to specify a coded block to be referred to when generating prediction values of DCT coefficients of a target block. For this reason, the prior art intra-frame prediction is not applied to the field DCT type macroblocks. It is therefore impossible to apply the intra-frame prediction to coding of the interlaced image or a specific progressive image in which the field DCT type macroblocks coexist. As a consequence, coding with high efficiency while reducing spatially redundant information satisfactorily is not realized.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide an image processing apparatus, and an image processing method wherein an interlaced image or a specific progressive image in which macroblocks of different DCT types coexist are coded efficiently while satisfactorily reducing spatially redundant image info
Burr & Brown
Kelley Chris
Matsushita Electric - Industrial Co., Ltd.
Philippe Gims
LandOfFree
Image processing method, image processing apparatus, 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 processing method, image processing apparatus, and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing method, image processing apparatus, and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2476584