Image analysis – Image compression or coding – Pyramid – hierarchy – or tree structure
Reexamination Certificate
1999-01-19
2002-12-31
Chen, Wenpeng (Department: 2624)
Image analysis
Image compression or coding
Pyramid, hierarchy, or tree structure
C382S239000, C375S240020, C375S240110, C375S240190
Reexamination Certificate
active
06501860
ABSTRACT:
The present invention concerns a digital signal coding device and method. It also concerns a decoding method and device corresponding to the coding method and device.
The purpose of coding is to compress the signal, which enables the digital signal to be transmitted, or respectively stored, thus reducing the transmission time or transmission rate, or respectively reducing the memory space used.
The invention is situated in the field of compression of digital signals with loss.
In a first aspect, breaking down a signal into frequency sub-bands before compressing it is known. The breakdown consists of creating, from the signal, a set of sub-bands which each contain a limited range of frequencies. The sub-bands can be of different resolutions, the resolution of a sub-band being the number of samples per unit length used for representing this sub-band. In the case of a digital image signal, a frequency sub-band of this signal can itself be considered to be an image, that is to say a bidimensional table of numerical values.
It should be noted that breaking down a signal into frequency sub-bands does not create any compression in itself, but makes it possible to decorrelate the signal so as to eliminate redundancy therefrom prior to the compression proper. The sub-bands are thus coded more effectively than the original signal.
A known method of coding a digital signal, in this case of a digital image, includes three main steps. The image is first of all broken down by a transformation into frequency sub-bands, then a scalar quantization of the coefficients thus obtained is performed. The quantized indices are finally coded by entropic coding without loss.
This method affords a high degree of compression of the signal. However, the ratio between the degree of compression and the coding error leaves room for improvement.
In the first aspect, the present invention aims to remedy the drawbacks of the prior art by providing a digital signal compression device and method which offers a high ratio of compression to distortion.
In second and third aspects, a known method of coding a digital signal, in this case of a digital image, is trellis coded quantization, which is described for example in the article entitled “Trellis Coded Quantization of Memoryless and Gauss-Markov Sources” by M. W. Marcellin and T. R. Fischer, which appeared in IEEE Transactions on Communications, Vol 38, Number 1, January 1990, as well as in the article “Universal Trellis Coded Quantization” by J. H. Kasner, M. W. Marcellin and B. R. Hunt, available by Internet at the address http://vail.ece.arizona.edu/Publications.html.
This method has the advantage of minimising the quantization error by virtue of the use of trellis coding.
In addition, in order to improve the ratio of compression to distortion, the inventors combined this method with another coding mode. In this case, blocks of the signal to be coded are allocated one or other coding mode, so as to select for each block the most appropriate coding mode, that is to say the one providing the highest ratio of compression to distortion.
Block here means a set of coefficients extracted from the signal in order to form a vector.
Coding by trellis coded quantization is applied to a series of blocks. For the coding to be optimal, it is necessary for the series to be fixed and known in advance. This means that, in order to allocate this mode to a given block in an optimum fashion, it would be necessary to know in advance all the blocks to which this code is allocated, and therefore to know the result of the allocation. There is therefore a difficulty in optimally allocating the coding modes to the blocks and consequently in obtaining the highest possible ratio of compression to distortion.
In the second aspect, the present invention aims to remedy the drawbacks of the prior art by providing a digital signal compression device and method which offers a high ratio of compression to distortion.
In the third aspect, the present invention aims to remedy the drawbacks of the prior art by providing a digital signal compression device and method which make it possible to allocate the trellis coded quantization coding mode to blocks formed in the signal, with great simplicity of implementation.
In a fourth aspect, the present invention aims to improve the first aspect of the invention, by applying a preprocessing between the step of breaking down into sub-bands and the coding step proper.
In the first aspect, the invention proposes a digital signal coding method, including the analysis of the signal in order to separate the pertinent information and the non-pertinent information, then the coding of the non-pertinent information according to a first coding mode which offers a high degree of compression, and coding of the pertinent information according to a second coding mode which limits distortion.
To this end, the invention proposes a digital signal coding method including a step of analysing the digital signal into a plurality of frequency sub-bands distributed in at least two different frequency bands, at least one first sub-band having a lower frequency and at least one second sub-band having a higher frequency,
characterised in that it includes, for each second sub-band, the steps of:
dividing the second sub-band into blocks,
selecting first blocks which are to be coded by setting to a predetermined value and second blocks which are to be coded by trellis coded quantization, according to a selection criterion,
linking the second blocks selected at the previous step in order to form a series of blocks,
coding the series of blocks by trellis coded quantization of a series of coefficients extracted from the second blocks of the second sub-band.
Correlatively, the invention proposes a digital signal coding device including means of analysing the digital signal into a plurality of frequency sub-bands distributed in at least two different frequency bands, at least one first sub-band having a lower frequency and at least one second sub-band having a higher frequency,
characterised in that it includes:
means of dividing each second sub-band into blocks,
means of selecting first blocks which are to be coded by setting to a predetermined value and second blocks which are to be coded by trellis coded quantization, according to a selection criterion,
means of linking the second blocks selected at the previous step in order to form a series of blocks,
means of coding the series of blocks by trellis coded quantization of a series of coefficients extracted from the second blocks of the second sub-band.
The breakdown into frequency sub-bands makes it possible to separate the pertinent information from the non-pertinent information. The non-pertinent information is then coded by setting to the predetermined value, which requires a nil transmission rate. Compression of the non-pertinent information is maximal.
It is thus possible to give a higher transmission rate to the pertinent information, present in the second blocks, which is thus coded more precisely. Distortion of the pertinent information is minimal, by virtue of the linking of the second blocks and their coding by trellis coded quantization.
Overall, the ratio of compression to distortion obtained is high.
The predetermined value is for example nil, so that, at decoding, blocks are constructed whose coefficients are nil.
According to a preferred characteristic, for each of the blocks the selection step includes the coding of the block by setting to the predetermined value and by a second coding mode, the comparison of the two coding modes according to the selection criterion and the selection of the coding by setting to the predetermined value if the block coded by setting to the predetermined value satisfies the selection criterion.
The selection is simple and rapid to implement. For example, the second coding mode is a scalar quantization of the coefficients of the block, or the second coding mode is a trellis coded quantization of a series of coefficients extracted from the blocks of the second sub-band. The latter case enabl
Charrier Maryline
Henry Felix
Onno Patrice
Canon Kabushiki Kaisha
Chen Wenpeng
Fitzpatrick ,Cella, Harper & Scinto
LandOfFree
Digital signal coding and decoding based on subbands does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Digital signal coding and decoding based on subbands, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Digital signal coding and decoding based on subbands will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2981299