Image analysis – Image compression or coding – Substantial processing of image in compressed form
Reexamination Certificate
2000-03-08
2004-01-27
Couso, Jose L. (Department: 2721)
Image analysis
Image compression or coding
Substantial processing of image in compressed form
C382S236000, C382S239000
Reexamination Certificate
active
06683987
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of Application
The present invention relates to a compressed data processing method and compressed data processing apparatus, and to a recording and playback system for compressed data, and in particular to such a method, apparatus and system for application to MPEG-encoded compressed data whereby a stream of pictures expressed by the data can be converted to an output picture stream having a reduced frequency of picture updating.
2. Description of Prior Art
Digital technology is widely applied at present in the fields of computers, broadcasting systems, communication systems, data storage systems, etc. A set of international standards for data compression known as MPEG (Moving Pictures Experts Group) has become an important part of such technology. Since the present invention utilizes the MPEG standards, the basic concepts of these will first be outlined. The first MPEG standards for compression of video data were developed by a joint committee known as ISO/IEC JTC1/SC2 of the ISO (International Standards Organization) and IEC (International Electrotechnical Commission) in 1988, where SC2 signifies “scientific sub-committee 2”, later changed to SC29.
There are two sets of MPEG standards, MPEG-1 and MPEG-2. MPEG-1 (signifying “MPEG phase 1”) is applicable to storage media etc., for transferring data at a rate of approximately 1.5 Mbps. MPEG-1 was developed by applying new technologies to existing types of picture encoding methods, specifically to the JPEG standard which is used for compression-encoding of still pictures, and the H.261 technology (specified by CCIT SGXV standards, now called the ITU-T SG15 standards), developed for compression of pictures in order to transmit the pictures at a low rate of data transfer in such applications as teleconferencing, video telephones, etc. with transmission via a ISDN network. The MPEG-1 standards were first published in 1993, as ISO/IEC 11172. MPEG-2 can be considered as an extension of MPEG-1, and was developed for applications such as data communications, broadcasting, etc., providing features which are not available with MPEG-1 such as an enhanced capability for compression encoding of interlaced-field video signals. The MPEG-2 standards were first published in 1994, as ISO/IEC 1318, H.262. Although embodiments of the invention will be described basically on the assumption of MPEG-1 (referred to in the following simply as MPEG) processing, it will be apparent to a skilled person that the techniques described can be readily adapted to MPEG-2 processing.
FIG. 19
is a general system block diagram showing an example of a basic configuration of an MPEG encoder. The operation will be described first for the case of forward prediction, i.e., deriving encoded data expressing a current input picture based upon the contents of a preceding reference picture, and considering only luminance values. In
FIG. 19
, data expressing successive ones of a stream of pictures are input to the decoder. Specifically, successive input pixel values of an input picture that is expressed as an array of pixels (for example, one frame of a progressive-scan video signal) are supplied to an adder
2
and a motion compensated prediction section
1
. The input picture is pre-processed (in some manner that is not indicated in the drawing) to extract successive 16×16 pixel blocks which are referred to as macroblocks, with the pixel values of the currently extracted macroblock being supplied to the adder
2
and motion compensated prediction section
1
. A picture memory
11
holds (as described hereinafter) a set of pixel values expressing a reference picture for use in processing a predictively encoded picture, or may hold a pair of reference pictures which respectively precede and succeed the a predictively encoded picture in the case of bidirectional encoding. With predictive encoding, the motion compensated prediction section
1
successively shifts the input macroblock with respect to the reference picture, within a predetermined search range, to determine whether there is a 16×16 array of pixels within the reference picture which has at least a predetermined minimum degree of correlation with the input macroblock. If such a condition is detected, then the amount and direction of displacement between that 16×16 pixel array in the reference picture and the input macroblock is obtained, as a vector value referred to as a motion vector (specifically, a combination of a horizontal and a vertical motion vector). The respective values of difference between the pixel values (i.e., luminance and chrominance values) of the input macroblock and the correspond pixels within that 16×16 array of pixels in the reference picture (read out from the picture memory
11
and supplied via the motion compensated prediction section
1
) are then derived by the adder
2
, and supplied to the DCT transform section
3
, with these values being referred to as motion compensated prediction error values in the following. Prediction from a preceding reference picture process is referred to as forward prediction, and from a succeeding reference picture is referred to as backward prediction. If no correlated 16×16 block is found within the search range, then the input macroblock is intra-coded within the input picture, i.e., as an intra-coded block, generally referred to as an I-block.
With bidirectional prediction, values for the input macroblock are predicted based on two 16×16 blocks of pixels within a preceding and a succeeding reference picture respectively.
In MPEG, the basic unit for which different types of encoding can be specified is the macroblock. Depending upon the type of picture in which it is located and on decisions made by the encoder, a macroblock may be:
(a) encoded entirely within a picture (i.e., intra-coded), independently of all other pictures,
(b) encoded by forward prediction, i.e., as a set of prediction error values in conjunction with a motion vector, derived using a preceding reference picture,
(b) encoded by backward prediction, i.e., as a set of prediction error values in conjunction with a motion vector, derived using a succeeding reference picture, or
(c) encoded by bidirectional prediction, using both a preceding and a succeeding reference picture.
A picture can be encoded as:
(a) an I-picture, in which case all of the macroblocks are I-macroblocks, i.e., are intra-coded within that picture,
(b) a P-picture, in which case the encoder can selectively apply intra-coding or forward prediction encoding to the macroblocks, or
(c) a B-picture, in which case the encoder can selectively apply intra-coding, forward prediction encoding, backward prediction encoding, or bidirectional prediction to the macroblocks.
To minimize the amount of generated encoded data, the encoder uses an algorithm which is designed to minimize the number of I-macroblocks of the P-pictures and B-pictures.
I-pictures and P-pictures are used as reference pictures, however B-pictures are not so used.
Successive ones of the stream of pictures supplied to the MPEG encoder are encoded as I, P or B-pictures, in a fixedly predetermined sequence. As a picture is encoded, the motion vectors derived for macroblocks are supplied from the motion compensated prediction section
1
to the VLC section
5
, as is also prediction mode information which specifies the macroblock type, i.e., whether that macroblock has been encoded by intra-coding, forward prediction, backward prediction, or bidirectional prediction.
The motion compensated prediction error values derived from the adder
2
for a macroblock of the input picture are supplied to a DCT transform section
3
, which processes the macroblock as a set of four 8×8 pixel blocks, sometimes referred to as DCT blocks. 2-dimensional DCT (Discrete Cosine Transform) processing is separately applied to each of these DCT blocks to obtain a corresponding set of DCT coefficients, which are supplied to a quantizer
4
. This form of processing is efficient, due to the fact that a video
Couso Jose L.
Dang Duy M.
Parkhurst & Wendel L.L.P.
Victor Company of Japan Ltd.
LandOfFree
Method and apparatus for altering the picture updating... 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 altering the picture updating..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for altering the picture updating... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3250225