1. Field of the Invention
The invention is related to video compression systems, and in particular to compression of digital video systems.
2. Discussion of the Background
Transmission of moving pictures in real-time is employed in several applications such as video conferencing, net meetings, TV broadcasting and video telephony.
However, representing moving pictures requires bulk information as digital video typically is described by representing each pixel in a picture with 8 bits (1 Byte). Such uncompressed video data results in large bit volumes, and can not be transferred over conventional communication networks and transmission lines in real time due to a limited bandwidth.
Thus, enabling real time video transmission requires a large extent of data compression. Data compression may, however, compromise the picture quality. Therefore, significant efforts have been made to develop compression techniques allowing real time transmission of high quality video over bandwidth limited data connections.
In video compression systems, a main goal is to represent the video information with as little capacity as possible. Capacity is defined with bits, either as a constant value or as bits/time unit. In both cases, a main goal is to reduce the number of bits.
The most common video coding method is described in the MPEG* and H.26* standards. The video data undergo four main processes before transmission, namely prediction, transformation, quantization and entropy coding.
The prediction process significantly reduces the amount of bits required for each picture in a video sequence to be transferred. It takes advantage of the similarity of parts of the sequence with other parts of the sequence. Because the predictor part is known to both the encoder and decoder, only the difference has to be transferred. This difference typically requires much less capacity for its representation. Further, the prediction is mainly based on picture content from previously reconstructed pictures where the location of the content is defined by motion vectors. The prediction process is typically performed on square block sizes (e.g., 16×16 pixels) as indicated in block M in
In a typical video sequence, the content of the present block M would be similar to a corresponding block in a previously decoded picture. If no changes have occurred since the previously decoded picture, the content of M would be equal to a block of the same location in the previously decoded picture. In other cases, an object in the picture may have been moved so that the content of M is more equal to a block of a different location in the previously decoded picture. Such movements are represented by motion vectors (V). As an example, a motion vector of (3;4) means that the content of M has moved 3 pixels to the left and 4 pixels upwards since the previously decoded picture. For improved accuracy, the vector may also include decimals, requiring interpolation between the pixels.
To reduce the data size of the motion vectors before transmission, it is assumed that the movements in one block are quite similar to the movements of the adjacent blocks. Thus, a prediction (Vpred) of V is created based on the actual motion vectors of the adjacent blocks. As the motion vectors of the adjacent blocks are already known at the receiving side, only the difference (Vdiff) between the actual motion vector and the corresponding prediction have to be represented: Vdiff=V−Vpred. At the receiving side, the motion vector could then be recreated by V=Vpred−Vdiff.
In the ITU standards H.261 and H.262 and the ISO standards MPEG1 and MPEG2, Vpred is set equal to the motion vector of the adjacent block corresponding to block a in
It is known from the publication US 2002/0039386 A1 a block matching processor and method for supporting block matching motion estimation at motion vector prediction modes using matching blocks of various sizes.
Further, it is known from US 2001/0031004 A1 a method and an apparatus for updating motion vector memories used for pre-diction of motion vectors within a video coding/decoding regime. The main issue in this document is how to store motion vectors in an efficient way for later use as current blocks. The prediction of the motion vectors is based on median calculation; hence this document does not describe a sufficiently accurate measurement method for motion vector prediction.
Still further there is known a solution (WO 01/99437 A2) where the prediction of motion vectors is based on median calculations. However, the main idea described in this document is to make a search within a smaller window in proximity to the predicted position.
The prediction vector derived according to the background art as described above has shown not to be sufficiently accurate. In addition, by selecting the prediction vector on a component by component basis, the vector may be constructed of components from different vectors resulting in a “fictional” motion vector.
Accordingly, one object of the present invention is to solve the above noted and other problems.
To achieve this and other objects, the present invention provides a novel method in video coding for determining a motion vector prediction associated with a target block of pixels representing a certain area of a video picture. The method includes the steps of calculating respective vector differences between a primary set of motion vectors or motion vector components respectively associated with a first number of blocks adjacent to the target block and one or more of a secondary set of motion vectors or motion vector components respectively associated with a second number of blocks also adjacent to the target block, and selecting a motion vector or motion vector component among said primary set of motion vectors corresponding to a smallest one of said vector differences as the motion vector prediction or a component of the motion vector prediction. The present invention also provides a novel computer program product.
A more complete appreciation of the invention and many of the attendant advantages thereof will be readily obtained as the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawing, wherein:
In the following, the present invention will be discussed by describing a preferred embodiment, and by referring to the accompanying drawing. However, a person skilled in the art will realize other applications and modifications within the scope of the invention as defined in the en-closed independent claim.
The present invention will now be described by an example, taking the notations of
It is further assumed that the motion vectors for a and b are known, and that one of them is chosen to be the prediction vector as a whole, even if a separate selection for each component is also a part of the present invention. Further, according to the present example, block c is selected to be decisive (the decision block) of the choice between a and b. C is selected because it is positioned in the right upper corner adjacent to block M, whereas a and b are positioned close to the opposite upper corner. Thus, the motion vector of c may provide a good indication of which motion vector is the best approximation to the motion vector of block M. Note that even if only one decision block is used here, several decision blocks could be used in the general case.
The block, a or b in this example, having the smallest vector difference related to motion vector of the decision block will preferably be selected as the motion vector prediction for M.
There are two preferred alternatives for calculating the vector difference. The idea behind the selection procedure for selecting the optimal motion vector prediction for M, is to select the block with the motion vector having the least vector difference compared to the motion vector of the decision block. Several methods apply when calculating these selection criteria. Two of them are shown in the following sections, both return a numerical value. This value will be used when deciding which block (a or c) to choose as the motion vector prediction for M, the smaller value is the best choice. The returned value will only have the function of a selection criterion, hence the value in itself does not have to be stored or transmitted. The first of the two alternatives calculates the squared sum of the difference of the respective components. Hence, the vector difference c−a and c−b would be calculated as follows:
√{square root over ((x3−x1)2+(y3−y1)2)}{square root over ((x3−x1)2+(y3−y1)2)} √{square root over ((x3−x2)2+(y3−y2)2)}{square root over ((x3−x2)2+(y3−y2)2)}
As long as it is only a question regarding which of the two returned values is the smallest for the expression above, one can choose to cancel the radical sign. The second alternative calculates the sum of the absolute values of the difference of the respective components. The corresponding calculation for the vector difference c−a and c−b would be as follows:
|x3−x1|+|y3−y1| |x3−x2|+|y3−y2|
In both cases, the motion vector having the smallest vector difference compared to the motion vector of c is selected as the prediction for the motion vector of block M.
In the general case, the number of vector candidates for the prediction is not limited to two as in the example above. In fact, any number of candidates can be used. Further, the number of decision vectors is neither limited to one. Thus, one set of vectors may be defined as the candidate set, while another set of vectors may be defined as the decision set.
This invention may be conveniently implemented using a conventional general purpose digital computer or microprocessor programmed according to the teachings of the present specification, as will be apparent to those skilled in the computer art. Appropriate software coding can readily be prepared by skilled programmers based on the teachings of the present disclosure, as will be apparent to those skilled in the software art. The invention may also be implemented by the preparation of application specific integrated circuits or by interconnecting an appropriate network of conventional component circuits, as will be readily apparent to those skilled in the art.
The present invention includes a computer program product which is a storage medium including instructions which can be used to program a computer to perform a process of the invention. The storage medium can include, but is not limited to, an type of disk including floppy disks, optical disks, CD-ROMs, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, magnetic or optical cards, or any type of pure software inventions (e.g., word processing, accounting, Internet related, etc.) media suitable for storing electronic instructions.
Obviously, numerous modifications and variations of the present invention are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims, the invention may be practiced otherwise than as specifically described herein.
Number | Date | Country | Kind |
---|---|---|---|
20025707 | Nov 2002 | NO | national |
Number | Name | Date | Kind |
---|---|---|---|
6122318 | Yamaguchi et al. | Sep 2000 | A |
20010004385 | Okumura et al. | Jun 2001 | A1 |
20010026590 | Kang et al. | Oct 2001 | A1 |
20010031004 | Kim et al. | Oct 2001 | A1 |
20020039386 | Han et al. | Apr 2002 | A1 |
Number | Date | Country |
---|---|---|
0 874 526 | Oct 1998 | EP |
WO0199437 | Dec 2001 | WO |
Number | Date | Country | |
---|---|---|---|
20040146110 A1 | Jul 2004 | US |