Image analysis – Image compression or coding – Including details of decompression
Reexamination Certificate
1998-10-23
2002-07-30
Tran, Phuoc (Department: 2621)
Image analysis
Image compression or coding
Including details of decompression
C382S236000
Reexamination Certificate
active
06427026
ABSTRACT:
BACKGROUND OF THE INVENTION
Field of the Invention
The invention relates to a video decoder and a corresponding method.
For coding video images, the MPEG standard has become established. It combines the principles of transformation and prediction coding with so-called hybrid coding. It contemplates three types of coded images, which are distinguished depending on the type of data compression:
1. I images (intra-coded images): The coding of an I image is done independently of other images. An I image is used to synchronize the transmission of a plurality of successive video images, and its coding utilizes the local correlation of the image data of the I image. The coding is preferably a discrete cosine transformation or DCT, followed by quantification and weighting of the transformation coefficients with ensuing entropy coding.
2. P images (predicted images): Their coding is dependent on a chronologically preceding I or P image (forward prediction). P images are subjected to motion estimation with respect to the preceding image (motion-compensated prediction). Next, intra-image coding (for example in the way described above for an I image) is used on the chronological prediction errors.
3. B images (bidirectionally predicted images): In these, chronological motion-compensated prediction is used, both with regard to a chronologically preceding and to a chronologically succeeding P or I image. The term “motion-compensated interpolation” is also used. The expressions (chronologically) “succeeding” or “preceding” refer not to the order of transmission of the images in the data stream of coded images but rather to their order of recording and playback. Images that are the basis for chronological prediction for coding another image are also known as support images. Like P images, B images are also coded in the form of quantified transformation coefficients of a differential image.
It is also known, particularly for showing on TV sets using the PAL or NTSC standard, for the playback of video images to be done not in the form of a so-called full frame but rather as two successive half-frames (“interlaced format”). A full frame has the total video image to be shown and thus includes all the lines of the image. Full frames are also called “frames”. Conversely, each of the two half frames (“fields”) includes only half of all the lines of one video image. The so-called “top field” or first half-frame involves all the even-numbered lines, while the “bottom field” or second half-frame involves all the odd-numbered lines of the associated full frame.
An MPEG decoder must accordingly decode not only the coded video images in the correct order but also, if they are to be shown on TV sets on the PAL or NTSC standard, must assure that the requisite two half-frames required for the purpose will be furnished. The video images transmitted may, however, be either frame-coded or field-coded. In the first case, the video images that are transmitted are full frames, while in the second case they are half-frames, two of each of which again form one full frame.
MPEG decoders have memory devices for storing previously decoded support images. Storing the I or P images in memory is necessary, because they are the basis for a chronological prediction of a P or B image and are thus needed for decoding of that image. If a coded B image received from the decoder is to be processed, for instance, then decoding of the B image is done with the aid of the two support images used for its bidirectional prediction. It is noted that the support images were previously received by the decoder and stored in memory.
It is also known to store the support images, required for decoding a B image, not in decoded form but rather in coded form in the decoder in order to save memory space. However, this method has the disadvantage that the data throughput to be handled by the decoder is sharply increased. The same I and P images in fact generally serve to decode many B images, so that they have to be often decoded correspondingly. An even more decisive factor, however, is that in MPEG, block-based coding is contemplated, and the prediction regions in the support images that are used for the prediction often deviate from the block limits. To decode one block of a B image, therefore, four blocks of each of the two coded memorized support images must therefore be decoded. This means eight times the data throughput, compared with storage of decoded support images, is needed. The eight times greater data quantity, however, must as before be handled in the same unit of time available in total for decoding the predicted coded image. The demands in terms of performance of the decoder are accordingly considerably greater than in the case of support images stored in decoded form.
In conventional MPEG decoders, a memory device is also provided in which the decoded B image is stored (“an output frame buffer”). The memory device is needed in conventional decoders in the case of a frame-coded B image, since as already noted two half-frames have to be created from the full frame by the decoder and if the full frame is to be shown in interlaced format, the two half-frames must be present successively at an output of the decoder. In the case of field-coded B images, however, the output frame buffer is still necessary for another reason. In some applications, it may be necessary (for instance for adaptation of different frame frequencies between the recording system and the playback system) for not merely two half-frames to be shown to make up one full frame but instead for the first half-frame to be repeated again after the second half-frame has been shown.
SUMMARY OF THE INVENTION
It is accordingly an object of the invention to provide a video decoder and a corresponding method, which overcomes the above-mentioned disadvantages of the prior art devices and methods of this general type, in which the memory requirement is low.
With the foregoing and other objects in view there is provided, in accordance with the invention, a video decoder, including: a memory storing a compression-coded image; a decoder having an output, the decoder receives the compression-coded image and performs a first decoding of the compression-code image and outputs a first decoded image at the output to be received by a playback device for playback; and the decoder subsequently performing a second decoding of the compression-coded image resulting in a second decoded image and outputting the second decoded image to the output.
According to the invention, the first image is stored in memory in the decoder not in decoded form (as in the prior art) but rather in coded form. To generate the necessary half-frames for playback, decoding is done at least twice. After each decoding operation, the outcome of decoding is output to an output of the decoder, in order to play it back (for instance using a TV set). An output frame buffer as in the prior art is not necessary. For storing the coded first image in memory, less memory space is necessary than if it were stored in decoded form as is the case in the prior art. The invention is not limited to the instance of MPEG decoders. It also relates to all cases where information must be generated multiple times from a first image for playback. Unlike the prior art, the stored, coded first image for generating the various information (such as two half-frames from one full frame) is newly decoded each time (each time, the image may need to be decoded only in part and not completely). In the prior art, complete decoding of the image is done first, the image is stored in memory and then output multiple times (for instance, outputting the two half-frames of a decoded full frame). The invention can also be used in decoding methods in which chronological prediction is not performed and no support images are provided.
In accordance with an added feature of the invention, there is a further memory connected to the decoder and stores a further image, the compression-coded image is predicted time-wise with respect to the further image in the decoder
Hein Soeren
Meier Wolfgang
Wehn Norbert
Greenberg Laurence A.
Lerner Herbert L.
Siemens Aktiengesellschaft
Stemer Werner H.
Tran Phuoc
LandOfFree
Video decoder and corresponding method 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 decoder and corresponding method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Video decoder and corresponding method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2841592