Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1994-03-14
2003-06-17
Rao, Andy (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240140
Reexamination Certificate
active
06580758
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method of coding images at very low data rate and a coding/decoding device implementing this method.
It applies especially to the production of low data-rate digital CODECs whose range of data rates lies between 8 and 16 Kb/s.
2. Discussion of the Background
In these coders the main coding modes used are intra-image coding modes and inter-image coding modes with movement compensation. The intra-image mode is, however, very little used due to the fact that, on the one hand, it is extremely costly in terms of number of bits transmitted, and that, on the other hand, sequence changes in the transmitted images generally pose enormous problems which are the cause, for example, of overflows of the storage capacity of buffer memories.
Among the known coding methods, those described, for example, by the “CITT specialist group on coding for visual telephony. Recommendation H261 version 11/1989”, standard is certainly the one which is best adapted for transmitting images at very low data rate since it operates in differential mode in the time domain and by cosine transform in the spatial domain. The major interest in working in differential inter-image mode is in reducing redundancy and thereby the quantity of information to be coded. The use of the mode of coding by cosine transform has the advantage of supplying a totally decorrelated image. However, it is necessary to use both direct cosine transforms and inverse cosine transforms since the predictions done within the coders and decoders take place in the spatial domain.
Moreover, that necessitates structuring the image into blocks and macroblocks, a macroblock consisting, for example, of four luminance blocks of 8×8 size and of two chrominance signal blocks. A 16×16 luminance macroblock represents, on the image, the same surface area as an 8×8 chrominance block by reason of the sub-sampling which is used. A low-pass filter is inserted into the coding loop in order to reduce the artificial effects in the high frequencies due to the movement compensation and to lower the quantification noise in the prediction loop. The choice of the coding mode in an a priori approach, before coding of the coefficients, is done chronologically by first of all making a choice between the mode compensated in movement and the inter-image mode, this choice being guided on measurements of inter-image differences, then a choice between the mode adopted and the intra-image mode by comparison of the activity, or variance, of the grey levels contained in the blocks to be transmitted.
An image memory containing the reconstructed preceding image makes it possible to operate in differential mode in the time domain via inter-image and movement-compensated inter-image predictors. According to the fill state of the buffer memory the coefficients are quantified then dequantified with a step which is calculated by the regulation stage. A variable-length coding provides for adapting of the code words to the statistical properties of the information to be transmitted. The role of the regulation is to make the various parameters for coding of the contents of the image to be transmitted evolve so as to ensure the required data rate. Video multiplexing has the function of generating a binary train containing, among other things, the coded coefficients, the intra, inter and modes compensated inter-image coding words and the synchronization words of the various structures of the images. However, although performing well, this regulation mode always involves the risk that an overflow of the buffer memory will take place, on initialization especially upon transmission of the first image or upon changing of the observed scene in image sequences, and forces transmission of the images to be stopped each time this risk is present, which is prejudicial to the integrity of the information transmitted.
The object of the invention is to alleviate the abovementioned drawback.
SUMMARY OF THE INVENTION
To this end, the subject of the invention is a method of differential coding of blocks of television images characterized in that it consists in coding the minimum pixel-by-pixel difference between the current block of the image to be transmitted and one among four blocks of the reconstructed current image in the closest vicinity of the current block, the closest block being the one which, among the four, gives the best prediction of the current block for the coding.
The subject of the invention is also a device for implementing the method.
The method and the device have the advantage that they make it possible to reduce the use of the intra-image coding mode, upon abrupt changing of an observed scene for example, upon transmission of the first image of a sequence or upon changing of scenes or variations in gains of the image sensor. The substitution of the coding mode according to the invention for the intra-image coding mode makes it possible considerably to reduce the cost of coding of the blocks and to avoid increasing the quantification step for the transformed coefficients. The invention also has the advantage that it makes it possible to reduce the time for convergence of the coder towards operation in steady-state regime.
REFERENCES:
patent: 4903128 (1990-02-01), Thoreau
patent: 4982285 (1991-01-01), Sugiyama
patent: 5063524 (1991-11-01), Ferre et al.
patent: 5093720 (1992-03-01), Krause et al.
patent: 5136377 (1992-08-01), Johnston et al.
patent: 5155594 (1992-10-01), Bernstein et al.
patent: 5164828 (1992-11-01), Tahara et al.
patent: 5198901 (1993-03-01), Lynch
patent: 5227878 (1993-07-01), Puri et al.
patent: 5260783 (1993-11-01), Dixit
patent: 5295201 (1994-03-01), Yokohama
patent: 5317397 (1994-05-01), Odaka et al.
patent: 5598216 (1997-01-01), Lee
patent: 0534282 (1993-03-01), None
Proceedings of the National Communications Forum, vol. 42, No. 2, Sep. 30, 1988, Oak Brook, Illinois, US, pp. 1678-1682. H. Gharavi “Low bit rate television on ISDN”.
Proceedings Pattern Recognition and Image Processing 1979, Chicago, Dec. 6, 1979, pp. 568-575, A. Ploysongsang Et Al. “DCT/DPCM processing of NTSC composite video signal”.
Signal Processing. Image Communication, vol. 1, No. 2, Oct. 1989, Amsterdam, Netherlands, pp. 103-115, K.Grotz et al. “A 64Kbit/s codec with forward analysis and control” voir p. 113, colonne de droite, ligne 14-ligne 19; figure 10.
CCITT—Recommendation H261, No. 584, Apr. 1990, pp. 120-128 “Codec for audiovisual services at nx384kbit/s” cite dans la demande voir le document en entier.
ICASSP 89 International Conference on Acoustics, Speech and Signal Processing, Glasgow; May 23, 1989, IEEE, New York, US; p. 1671-1674, Saito et al, “Self organizing pattern-matching coding for picture signals” voir abrege; figures 1,2; voir p. 1672, colonne de gauche, ligne 13—colonne de droite, ligne 6.
Chevance Christophe
Thoreau Dominique
"Thomson-CSF"
Oblon & Spivak, McClelland, Maier & Neustadt P.C.
Rao Andy
LandOfFree
Method of coding images at very low data rate and... 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 of coding images at very low data rate and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of coding images at very low data rate and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3099684