Image analysis – Image compression or coding – Pyramid – hierarchy – or tree structure
Reexamination Certificate
2000-06-15
2004-03-23
Couso, Jose L. (Department: 2621)
Image analysis
Image compression or coding
Pyramid, hierarchy, or tree structure
Reexamination Certificate
active
06711298
ABSTRACT:
The present invention concerns, in general terms, digital signal coding and proposes to this end a device and method for coding a digital signal by decomposition into frequency sub-bands of the signal, and segmentation of certain frequency sub-bands. It also concerns the transmission of the coded signal and also a decoding method and device corresponding to the coding device and method.
The purpose of coding is to compress the signal, thus making it possible to transmit the digital signal or to store it by reducing the quantity of binary symbols necessary for representing it. The coding can be without loss, that is to say it keeps all the information contained in the digital signal, or on the other hand with loss, that is to say some information contained in the digital signal may be degraded.
The present invention is applicable in each of the above two types of digital signal coding. Hereinafter, the coding of digital images or video sequences will be dealt with more particularly. A video sequence is defined as a succession of digital images. It is particularly adapted to the storage of images in data bases and to their transmission over a network to a number of distant items of equipment.
It is known that a digital signal can be decomposed into frequency sub-bands before compressing it. Decomposition consists of creating, from the digital signal, a set of sub-bands each containing a limited frequency spectrum. 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 be considered to be an image, that is to say a bidimensional table of digital values.
It should be noted that the decomposition of a signal into frequency sub-bands creates no compression in itself, but makes it possible to decorrelate the signal so as to eliminate the redundancy existing in the digital image prior to the compression proper. The sub-bands are then compressed more effectively than the original signal.
Conventionally, the coding of a digital signal, in this case of a digital image, includes three steps. The image is first of all decomposed by a transformation into frequency sub-bands, the coefficients thereof are quantised as indices and finally these indices are coded by means of an entropic coding without loss.
This type of compression makes it possible to obtain a relatively high degree of compression of the signal but does not make it possible to access the content of the image. In other words, the decomposition of the signal remains purely of the frequency type, and gives no information about the objects which may be contained in the image. Object means an entity of the image corresponding to a semantic unit, for example the face of a person. An object can comprises one or several regions of the image. In the following, notions of object and region will be considered as equivalent.
Coding using a decomposition into sub-bands of the signal is by nature progressive by sub-band, and therefore allows transmission of the coded data which is progressive by sub-band.
There also exist other image compression techniques based on the segmentation thereof. In this context, the image is considered to consist of objects with two dimensions. Segmentation is a low-level process whose purpose is to effect a partitioning of the image into a certain number of subelements referred to as regions. The partitioning is such that the regions are separate and combining them forms the image. The regions correspond or do not correspond to objects in the image, the term object referring to an item of information of a semantic nature. Very often, however, an object corresponds to a region or set of regions. Each region can be represented by an item of information representing its shape, colour or texture.
Conventionally, a method of compressing a digital image based on a segmentation includes a first so-called marking step, that is to say the interior of the regions having local homogeneity is extracted from the image. Next, a decision step precisely defines the contours of the areas containing homogeneous data At the end of this step, each pixel of the image is associated with a label identifying the region to which it belongs. The set of all the labels of all the pixels is conventionally referred to as a segmentation map. Finally, in such a coding, the last step consists of coding the segmentation map, generally in the form of contours of the regions, and pertinent parameters representing the interior of the regions, such as the texture and the colour.
This type of technique makes it possible, for a given image, to obtain a higher degree of compression than with the technique previously described. This is because, with segmentation, the compression can be effected selectively on the object or regions judged to be the most important, to the detriment of the others. Thus, for a given degree of compression, that is to say for a number of binary elements allowed, a precise object (typically the face of a person in an image of the “head and shoulders”) type, can be coded precisely using a maximum number of bits, to the detriment of the background, which for its part will be coded with a minimum number of bits.
Segmentation allows progressive coding by regions, and consequently transmission of the coded data which is progressive by regions.
This type of technique, however, does not make it possible to have multiresolution information as permitted by the methods with decomposition into sub-bands.
Other techniques combine the two compression methods described above, such as for example the standard known as MPEG4 (from “Motion Picture Expert Group”), which is currently being converted into an 30 ISO/IEC standard. In the MPEG4 coder, more particularly in the case of the coding of fixed images, the decomposition of the image into frequency sub-bands is used conjointly with a segmentation of the image. A step prior to the coder (not standardised) is responsible for isolating the objects in the image (Video objects) and representing each of these objects by a mask. In the case of a binary mask, the spatial support of the mask has the same size as the original image and a point on the mask at the value 1 (or respectively 0) indicates that the pixel at the same position in the image belongs to the object (or respectively is outside the object).
For each object, the mask is then transmitted to a shape decoder whilst the texture for each object is decomposed into sub-bands, and the sub-bands are next transmitted to a texture decoder.
This method has a certain number of drawbacks. This is because it is necessary to code the mask for each object at its highest resolution level (in Version 1 of MPEG4) or in certain cases at two resolution levels (Version 2 of MPEG4). For a given degree of compression, this impairs the quality of the reconstructed image, since it is necessary to reserve output for the masks of the objects, which are at a high resolution. Moreover, the number of objects handled is a priori the same at all levels, whilst it may be more advantageous to have a number of objects increasing with the (spatial) resolution, that is to say a true conjoint scalability between the resolution and the number of objects.
The present invention aims to remedy the drawbacks of the prior art, by providing a method and device for compressing a digital signal which offer a high compression ratio whilst allowing progressive transmission of the content of the image, both in resolution and by objects.
To this end, the invention proposes a method of coding a set of data representing physical quantities, characterised in that it includes the steps of:
decomposing the set of data into a plurality of frequency sub-bands on at least one resolution level,
coding the sub-bands,
then, for each resolution level,
segmenting at least one sub-band into at least two homogeneous regions, in order to form a segmentation map,
ordering the regions according to a predetermined criterion,
Canon Research Centre France S.A.
Couso Jose L.
LandOfFree
Digital signal coding, decoding and transmission, with... 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, decoding and transmission, with..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Digital signal coding, decoding and transmission, with... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3200866