Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
2000-03-23
2002-07-09
Diep, Nhon (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240230
Reexamination Certificate
active
06418168
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a motion vector detection apparatus and method of the same and to an image processing apparatus.
2. Description of the Related Art
For example, the apparatus shown in
FIG. 12
has been known as an image processing apparatus performing image compression and coding.
A pre-processing circuit
102
in an image processing apparatus
101
performs processing for conversion of the image format of image data S
100
and outputs image data S
102
a
obtained by the conversion processing to an image memory
103
and a motion detection circuit
104
.
The motion detection circuit
104
, as shown in
FIG. 13
, performs motion detection between a prior frame
150
input from the image memory
103
and a present frame
151
input from the pre-processing circuit
102
in units of macroblocks MB of 16×16 pixels within a search area
152
indicated by a search area designation signal S
108
a
input from a control circuit
108
to produce a motion vector
153
.
The reason for performing the motion detection on a search area defined in part of an effective screen is that if the motion detection is performed over all of the effective screen, the amount of calculation becomes enormous, the apparatus becomes large in scale, and the processing time becomes long.
The produced motion vector or a residual vector produced using the motion vector is subjected to DCT processing in a DCT/quantization circuit
105
to produce a DCT coefficient. The DCT/quantization circuit
105
quantizes the DCT coefficient using a quantization step indicated by a Q-scale signal S
108
b
input from the control circuit
108
, then outputs it to a variable length coding circuit
106
.
The quantized DCT coefficient is coded by variable length in the variable length coding circuit
106
, then is output through a buffer memory
107
to a recording system or a transmission system.
A method of determining the search area of motion detection by the motion detection circuit
104
by a conventional control circuit
108
will be explained below.
For example in a Moving Picture Experts Group (MPEG) and other moving picture compression systems, in order to compress a series of frames constituting a moving picture using bidirectional prediction, coding is performed by three types of pictures, that is, an intra-coded (I) picture, a predictive-coded (P) picture, and a bidirectional (B) picture. An I-picture is a frame which is obtained by coding an input frame as it is without prediction, a P-picture is a frame using one-directional motion compensation prediction, and a B-picture is a frame using bidirectional prediction.
Within image data S
100
, as shown in
FIG. 14
, pictures located in time in the order of B
0
, P
1
, B
2
, B
3
, and P
4
pictures are encoded in the image processing apparatus
100
in the order of the B
0
, P
1
, P
4
, B
2
, and B
3
pictures on the basis of the relationship of previous and following pictures.
Here, B and P indicate the B-picture and P-picture respectively. The numbers following them indicate the order of location in time in the image data.
Conventionally, when the B
0
picture is encoded, a motion vector frmvBP from a B
0
picture to P
2
picture is found in units of macroblocks. Next, when the P
4
picture is encoded, since the distance in time between the P
1
picture and the P
4
picture is three times as long as that between the B
0
picture and the P
1
picture, assuming that the same motion as the motion vector frmvBP continues, as shown in
FIG. 15
, a motion vector frmvPP from the P
1
picture to the P
4
picture is found by tripling the motion vector frmvBP.
In the conventional method described above, for example, as shown in
FIG. 16A
, when the motion vector frmvBP, the motion vector frmv between the P
1
picture and the B
2
picture, and the motion vector frmv
3
between the B
3
picture and the P
4
picture are almost the same, that is, when almost the same motion continues, a suitable search area can be set.
Summarizing the problem to be solved by the invention, the method of the related art described above suffered from the problem that when there is a difference between the motion vector frmvBP, motion vector frmv
1
, motion vector frmv
2
, and motion vector frmv
3
, that is, when the motion of the image is severe (rapid), the motion vector frmvPP found by tripling the motion vector frmvBP and the actual motion vector obtained by adding the motion vectors frmv
1
, frmv
2
, and frmv
3
no longer match, a suitable search area can no longer be set.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a motion vector detection apparatus and method of the same and an image processing apparatus able to determine a suitable search area when detecting a motion vector for a rapid motion image.
According to a first aspect of the present invention, there is provided a motion vector detection apparatus for detecting a motion vector of image data between a plurality of frames, comprising a motion detection circuit specifying image data of a first block matching or almost matching image data of a first block within a first frame in a search area designated within a second frame located before the first frame in time and detecting a motion vector from a positional relationship between the first block of the first frame and the specified first block in the search area of the second frame; a first control circuit specifying the search area by using motion vector of image data found in units of the first blocks between a plurality of frames located before the second frame in time; a simplified motion vector detection circuit detecting a simplified motion vector indicating motion of an image in units of second blocks larger than the first block by using the first frame, the second frame, and a frame located between the first frame and the second frame; a second control circuit specifying the search area by using the simplified motion vector; and a third control circuit judging whether or not a change of motion between the first frame and the second frame is larger than a predetermined standard, designating a search area specified by the second control circuit in the motion detection circuit when the change of motion is larger than the predetermined standard, and designating a search area specified by the first control circuit in the motion detection circuit when the change of motion is not larger than the predetermined standard.
Preferably, the third control circuit finds a cumulative sum of differences which is defined as a cumulative sum of absolute values of differences between simplified motion vectors adjacently located in time within a predetermined section, judges whether or not the cumulative sum of differences is larger than a predetermined standard value, designates the search area specified by the second control circuit in the motion detection circuit when judging the cumulative sum of differences is larger than the standard value, and designates the search area specified by the first control circuit in the motion detection circuit when judging the cumulative sum of differences is not larger than the standard value.
Preferably, the standard value is a fixed multiple of an average value of magnitude of the simplified motion vector detected within the predetermined section.
Alternatively, preferably, the third control circuit finds the cumulative sum of differences by removing from the detected simplified motion vectors in a predetermined section a simplified motion vector having a difference more than a predetermined standard value between average values of two simplified motion vectors located before and after each other adjacent in time.
Preferably, the first control circuit finds a motion vector between the second frame and a third frame located immediately before in time from the second frame in units of the first blocks and specifies the search area by using the motion vector which is found by multiplying the found motion vector by a number defined by a distance of time between
Diep Nhon
Frommer William S.
Frommer & Lawrence & Haug LLP
Sony Corporation
LandOfFree
Motion vector detection apparatus, method of the same, 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 Motion vector detection apparatus, method of the same, and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Motion vector detection apparatus, method of the same, and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2850880