Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-11-17
2004-12-07
Kelley, Chris (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240100, C375S240010
Reexamination Certificate
active
06829303
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to decoding of images and, more specifically, to methods and apparatus for performing video decoding in a system including a programmable processor, e.g., a central processing unit (CPU).
BACKGROUND OF THE INVENTION
The digital represent ation of images is becoming ever more common. Digital video, e.g., digital high definition television (HDTV), represents one application where digital video signals comprise, e.g., a sequence of digital images. With the increasing use of the Internet, the generation, storage and transmission of digital video images is likely to continue to increase.
In order to reduce the amount of data required to represent digital images, such images are often encoded, e.g., compressed. Coding of images in the form of frames may be performed on an intra-frame basis so that the generated encoded image data does not depend on data from other images. Intra-frame coding is often performed using discrete cosine transform (DCT) coding techniques. Data generated from intra-frame coding is often called intra-coded data.
When a sequence of images is being coded, coding of images may be performed on an inter-frame, as well as an intra-frame, basis. Inter-frame coding generally involves using motion compensated prediction techniques to produce motion vectors which include information on how a portion of a subsequent image may be generated using a portion of a preceding or subsequent image as reference data. Motion vector information, once generated, may be coded, e.g., using differential coding techniques, to reduce the amount of data required to represent the motion vector information.
When there is little motion between frames, motion vectors provide an efficient method of representing portions of image. However, intra-frame coding can be more efficient when there is a substantial change from one frame to another. In an attempt to maximize coding efficiency, often some portions of an inter-coded frame are encoded using intra-frame coding techniques while other portions of the same image are coded using inter-frame coding, e.g., motion vectors. Those portions of inter-coded frames that use inter-frame coding, may also include coded information regarding the residual, or correction, image. During decoding of such portions of frames, the coded residual data is decoded. The decoded residual image data is combined with image data generated by performing motion compensated predictions using reference frame data and motion vectors. Those portions of inter-coded frames that use intra-frame coding are decoded without the use of reference frame data or motion vectors. In this manner, a complete inter-coded frame can be generated from the intra- and inter-coded data used to represent the frame.
MPEG-2 is a well known video standard which includes support for the use of motion compensated prediction techniques in addition to transform coding, e.g., the use of DCT transforms. The international video standard MPEG-2 is described in the International Standards Organization document (ISO/IEC 13818-2). MPEG2 has been used as the basis for several commercial applications including digital video disks (DVD) and digital broadcast television.
In MPEG-2 motion vector information is differentially encoded for transmission purposes. In addition, quantization operations and run length coding operations are performed on video data to further reduce the amount of data required to represent the images being encoded. A scan conversion operation is also normally performed as part of the MPEG-2 coding process in order to convert a set of two dimensional coefficient data into a one dimensional data sequence which can be processed, stored and/or transmitted. Variable length coding is applied to many of the data elements in order to further reduce the number of bits needed to represent an image.
In order to view images represented using encoded video data, the image data has to be decoded prior to display. MPEG-2 decoding generally involves performing operations which are the inverse to those used to originally encoded the image data, e.g., decoding usually involves variable length decoding, inverse scan conversion, inverse quantization, inverse discrete cosine transform (IDCT), motion vector reconstruction and motion compensated prediction operations.
Given the large amount of data used to represent images, a considerable amount of processing is normally required to decode an encoded image. In the case of motion vectors, image decoding is complicated by the need to access reference frame data, e.g., previously decoded frames, in order to generate a current decoded frame. The need to access reference frame data to perform motion compensated predictions using motion vector information results in the motion compensated prediction operations being highly memory intensive. The large number of memory access operations associated with performing motion compensated predictions can often have a significant impact on the amount of resources required to decode an image.
Users of computers are beginning to expect that they will be able to decode and display video images, e.g., encoded motion pictures, in real time. In order to achieve the real time display of encoded video, the encoded video images, e.g., frames, on average, need to be decoded in the same or a smaller amount of time than is used to display the images. As discussed above, the amount of processing required to decode an image can be considerable. Placing real time restraints on decoder circuitry further complicates matters due to the time constraints in which the decoding must be accomplished.
Generally, known video decoders belong to one of three types: 1) those that use dedicated special purpose integrated decoder circuits coupled with video memory to fully decode encoded video data; 2) those that use software and a general purpose programmable processor, e.g., CPU such as a Pentium processor, to fully decode encoded video data; and 3) those that partition the video decoding between a general purpose programmable processing unit such as a Pentium processor and a graphics processor chip which is used to manage and/or perform memory intensive operations such as motion compensated prediction. A high speed video or graphics memory is often used in conjunction with the graphic's processor to further accelerate image decoding operations.
In many high definition television sets and other video display devices, the first type of known decoders are used. That is, dedicated hardware decoder circuits are used to decode received encoded video data.
FIG. 1
illustrates a known decoder circuit
100
capable of decoding MPEG-2 video signals.
As illustrated, the video decoder
100
comprises a variable length decoder (VLD) circuit
102
, an inverse scan circuit
104
, an inverse quantization circuit
106
, an inverse discrete cosine transform (IDCT) circuit
108
, a motion compensated prediction circuit
110
, and a frame store memory
112
. Motion vector reconstruction circuitry, which is used to reconstruct motion vectors from encoded motion vector information, is present but not explicitly shown in the
FIG. 1
illustration.
In the known system, the VLD circuit
102
receives MPEG-2 encoded video data and performs a variable length decoding operation on variable length encoded data included therein. The inverse scan circuit
104
is responsible for re-sequencing elements of the video data output by the VLD circuit
102
to reverse the effect of the scan conversion operation originally used to convert the two dimensional coefficient data into a one dimensional ordering. The inverse quantization (IQUANT) circuit
106
performs an inverse quantization operation on the quantized values included in the video data output by the inverse scan circuit
104
. The IQUANT circuit
106
generates a video data stream including DCT coefficients. The DCT coefficients are processed by the IDCT circuit
108
to generate decoded image date from received intra-coded image data. The output of the IDCT circuit, in the ca
Pearlstein Larry
Sazzad Sharif Mohammad
Hitachi America Ltd.
Kelley Chris
Parsons Charles E
Straub Michael P.
Straub & Pokotylo
LandOfFree
Methods and apparatus for decoding images using dedicated... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods and apparatus for decoding images using dedicated..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods and apparatus for decoding images using dedicated... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3281271