Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Utility Patent
1998-08-25
2001-01-02
Le, Vu (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C348S699000
Utility Patent
active
06169766
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to technology for efficiently detecting a motion vector during the motion compensated coding of a moving picture.
A block matching method is generally used for detecting a motion vector out of a moving picture.
FIG. 8
is a conceptual diagram illustrating a block matching method. As shown in
FIG. 8
, according to the block matching method, an estimated value representing correlation is first calculated between an image component included in a block to be coded, out of which the motion vector is to be detected, and an image component in each candidate block within a search range, which block has the same size as the block to be coded. Then, the displacement from a block, located at the same position as the block to be coded, to a candidate block having the highest correlation represented by the estimated value is defined as the motion vector. As the estimated value of correlation, a total sum of the absolute values of differences between pairs of mutually associated pixels in the pair of blocks is used, for example. In this case, the higher the correlation between the image components (i.e., pixels) is, the smaller the estimated value becomes.
In the block matching method, the estimated values are calculated between a single block to be coded and many candidate blocks within the search range. Thus, in order to obtain the numerous estimated values, a huge quantity of calculation is required. As a result, the power consumed by a motion vector detecting apparatus is also enormous.
In detecting a motion vector in accordance with the block matching method, various techniques are employed for reducing the quantity of estimated value calculation and thereby reduce the power consumption. Sub-sampling (decimation) is one of such techniques. In accordance with the sub-sampling method, some pixels of a block to be coded and the counterparts of the blocks within the search range are decimated, and a motion vector is obtained with respect to the decimated images by the block matching method. In the sub-sampling method, if the number of pixels included in the block to be coded is reduced to 1/N (where N is a positive integer) through decimation, then the quantity of estimated value calculation is also reduced to 1/N. As a result, the power consumption can also be reduced to about 1/N.
However, since some pixels are decimated from the block to be coded and the respective blocks within the search range in accordance with the sub-sampling method, it is probable that the decimated image loses the detailed features of the original image. Thus, as compared with a case of not using sub-sampling, the quantity of estimated value calculation can be reduced, but the performance of motion vector detection is inferior. Nevertheless, as the case may be, the reduction of the quantity of estimated value calculation and the power consumption is preferred even if the performance of motion vector detection is deteriorated to a certain degree.
SUMMARY OF THE INVENTION
The object of the present invention is selectively and more adaptively setting sub-sampling than conventional technology in terms of detection performance and power consumption.
Specifically, the present invention is embodied in a motion vector detecting apparatus. The motion vector detecting apparatus includes: first image storage means for storing an image of a block to be coded; second image storage means for storing an image within a search range; correlation calculation means including a plurality of correlation calculation blocks for calculating correlation between the block to be coded and each associated candidate block within the search range, each said correlation calculation block being supplied with a different group of pixels out of the block to be coded from the first image storage means and an associated group of pixels of each said associated candidate block from the second image storage means and calculating the correlation between the two groups of pixels supplied; control means for controlling the correlation calculation means by selectively operating at least one of the correlation calculation blocks; and correlation determination means for determining the correlation between the block to be coded and each said associated candidate block based on an output of the correlation calculation block operated in accordance with the control of the control means.
The control means preferably selects at least one of the correlation calculation blocks in accordance with image feature information of the block to be coded or operating state information of the motion vector detecting apparatus.
The present invention is also embodied in a method for detecting a motion vector. The method includes the steps of: a) calculating correlation between a block to be coded in a target frame and each associated candidate block within a search range in a search frame; and b) detecting the motion vector of the block to be coded based on the correlation calculated in the step a). In the step a), sub-sampling during the calculation of correlation is selectively set in accordance with image feature information of the block to be coded.
The frequency components of the block to be coded are preferably used as the image feature information. Alternatively, the DCT components or the motion vector of a block in a frame previous to the target frame may also be used as the image feature information. The block in the previous frame is preferably located at the same position as or in the vicinity of the block to be coded.
Another method for detecting a motion vector according to the present invention includes the steps of: a) calculating correlation between a block to be coded in a target frame and each associated candidate block within a search range in a search frame; and b) detecting the motion vector of the block to be coded based on the correlation calculated in the step a). In the step a), sub-sampling during the calculation of correlation is selectively set in accordance with operating state information of an apparatus implementing the method.
Information indicating whether the apparatus implementing the method is operated in a mobile state or in an immobile state is preferably used as the operating state information. Alternatively, information indicating the power left in a battery functioning as a power supply for the apparatus implementing the method may also be used as the operating state information.
The present invention is further embodied in an image coding apparatus for performing motion compensated coding on a moving picture. The image coding apparatus includes a motion vector detecting section for calculating correlation between a block to be coded in a target frame and each associated candidate block within a search range in a search frame and thereby detecting the motion vector of the block to be coded based on the correlation calculated. The motion vector detecting section selectively sets sub-sampling during the calculation of the correlation in accordance with image feature information of the block to be coded.
The image coding apparatus preferably further includes a DCT section for performing a discrete cosine transform on an image. The motion vector detecting section preferably selectively sets sub-sampling by using DCT components of a block in a frame previous to the target frame as the image feature information. The DCT components are obtained by the DCT section, and the block in the previous frame is located at the same position as or in the vicinity of the block to be coded.
The motion vector detecting section preferably selectively sets sub-sampling by using a motion vector of a block in a frame previous to the target frame as the image feature information. The block in the previous frame is located at the same position as or in the vicinity of the block to be coded.
REFERENCES:
patent: 5712799 (1998-01-01), Framwald et al.
patent: 5731850 (1998-03-01), Maturi et al.
patent: 5825426 (1998-10-01), Wickstrom et al.
patent: 0 732 670 (1996-09-01), None
patent: 63296583 (1988-12-01
Aoki Katsuji
Araki Toshiyuki
Ohtani Akihiko
Le Vu
Matsushita Electric - Industrial Co., Ltd.
McDermott & Will & Emery
LandOfFree
Method and apparatus for detecting motion vector and image... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for detecting motion vector and image..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for detecting motion vector and image... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2469212