Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-10-15
2001-08-07
Kelley, Chris (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240160, C382S232000, C382S240000, C382S246000
Reexamination Certificate
active
06272178
ABSTRACT:
The present invention relates to an encoder for performing motion compensated encoding of video data as well as to a decoder for decoding motion compensation encoded video data. The present invention furthermore relates to a method of motion compensated encoding of video data and a method of decoding motion compensation encoded video data.
Motion compensating (MC) video encoders and decoders utilize the fact that consecutive frames of a typical video sequence show a high degree of similarity with each other and the fact that the difference between consecutive frames is mostly caused by motion in the scene. In this way, motion compensating video encoding enables compressing the amount of video data as is known for instance from Proceedings of IEEE, volume 73, No. 4, pages 523-548, April 1995, H. Mussiuann, P. Pirsch und H. Grallert: “Advances in Picture Coding”. Instead of compressing each frame I(x,y) of a video sequence separately, MC video encoders compress the difference between a current video frame which has to be coded and a so called prediction frame P(x,y) which is calculated in the encoder and represents an estimate of the current video frame I(x,y) based on one or more of the previous video frames. The compressed difference between the current video frame I(x,y) and the prediction frame P(x,y) can be used by a video decoder for reconstructing the current video frame I*(x,y) by means of performing a similar prediction of the frame as performed by the encoder and by means of correcting this prediction frame based on the received difference between the prediction frame P(x, y) and the actual frame I(x, y).
It is common in the video coding art that different areas of the image are coded using different coding modes. This is the case in all modern video codecs such as the ITU H.261 and H.263 as well as the MPEG-1 and MPEG-2 video coding standards. For example, some of the image areas are coded without using any temporal prediction (so-called intra-blocks). Therefore, the term “frame” stands in the following for the areas of the frame subject to motion compensated encoding and decoding. It is to be noted that with the term “frame” in the context of motion compensated video coding and decoding no limitation to the specific case that no intrablocks are present, is intended.
Both the encoder as well as the decoder base the prediction of the current frame I(x,y) on estimated motion vectors which are calculated in the encoder by means of motion estimation. The estimated motion vectors are used in the encoder for constructing the prediction frame P based on which the encoder can calculate the prediction error frame E(x,y), i.e. the difference between the current frame I and the prediction frame. Moreover, the prediction error frame E(x,y) and the estimated motion vectors describing estimated pixel displacements of the current frame I relative to the reference frame R are communicated to the decoder for prediction of the current frame P based on the motion vectors and correction of the prediction frame based on the communicated prediction error, thus constructing the current frame I*(x,y).
In order to further reduce the amount of data communicated to the decoder, the encoder performs compression of the prediction error. As a result of compression of the prediction error and decompression of the compressed prediction error, the decompressed prediction error frame E*(x,y) may be slightly different from E(x,y). The encoder takes account of this slight deviation by means of basing motion estimation and calculation of the prediction frame P(x,y) as well as calculation of the prediction error frame E(x,y) not on preceding original frames I but rather on preceding frames I* reconstructed by the encoder in a way similar to the reconstruction performed by the decoder.
For motion estimation both reconstructed frames I* or original frames I can be used.
The compression of the prediction error frame E
n
(x,y) is typically achieved by partioning the whole frame into smaller areas and by applying an appropriate compression method to each of these areas. The term “area” stands for a number of pixels forming a contiguous spatial two dimensional block in a frame. The prediction error is typically concentrated in only few areas whereas in the remaining areas it is very small or zero. Accordingly, the compressed prediction error data conventionally comprise two types of information, namely addressing information specifying the areas of the prediction error frame which are coded, and furthermore the values of the compressed prediction error in these areas.
From Proc. of 1992 European Conference on Signalling Processing EUSIPCO, pages 1357-1360, N. Noresco, F. Lavagetto and F. Cocurullo, “Motion Adaptive Vector Quantization for Video Coding” an algorithm for motion adaptive video coding is known according to which video input frames are subdivided into three dimensional (spatio-temporal) blocks of fixed size and are segmented depending on the motion activity in each block. Depending on the motion content, the blocks are subsampled in time by skipping a certain number of frames. The 2 dimensional spatial slices of the block which have not been skipped, are vector quantized and coded. At the receiver the spatial slices of each block which have been transmitted are decoded and the slices in between are linearly interpolated leading to the reconstruction of the whole block.
It is the object of the present invention to improve encoding and decoding of video data such that higher encoding efficiency can be achieved and the bit rate of the encoded video data can be further reduced.
According to the present invention, this object is solved by an encoder for performing motion compensated encoding of video data according to claim
1
and furthermore by a decoder for decoding motion compensation encoded video data according to claim
15
. Moreover, this object is solved by a method of motion compensated encoding of video data according to claim
32
and also by a method of decoding motion compensation encoded video data according to claim
33
. Embodiments of the invention are defined in the dependent claims.
According to a first aspect of the present invention, properties of the prediction frame and/or properties of the motion vectors are utilized for reducing the bit rate of coded video data and hence, for improving the coding efficiency of video data. Prediction error concentrates along the contours of moving objects in a video scene. The invention is based on the observation that structures in the prediction frame P(x,y), i.e. spatial variations or patterns of luminance and/or chrominance pixel values, e.g. edges and the like, approximately coincide with locations of these contours of moving objects. Throughout this document the term “discontinuity of image brightness” denotes such general structures of the prediction frame. The prediction frame P(x,y) in turn is known both to the encoder and the decoder as soon as motion vectors are transmitted. Accordingly, based on the prediction frame it is possible to determine both in the encoder and in the decoder, which areas in the prediction error frame are significant and are to be encoded and decoded, respectively, without the need of transmitting address information regarding such areas of the prediction error frame contained in the encoded video data.
The present invention is furthermore based on the recognition that the prediction error concentrates in areas of high motion activity. Accordingly, properties of the motion vectors available both in the encoder as well as in the decoder can be used for determining significant areas of the prediction error frame which are to be encoded by the encoder and which are to be decoded by the decoder, without the need of including address information in the coded video data regarding the significant areas of the prediction error frame.
A preferable video data coding scheme according to this aspect of the present invention makes use both of structural characteristics of the prediction frame available both in the encoder a
Haavisto Petri
Kalevo Ossi
Nieweglowski Jacek
Tûrker Mustafa Ali
An Shawn S.
Kelley Chris
Nokia Mobile Phones Ltd.
Perman & Green LLP
LandOfFree
Video data encoder and decoder does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Video data encoder and decoder, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Video data encoder and decoder will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2477836