Image analysis – Image compression or coding
Reexamination Certificate
1999-05-24
2003-10-21
Johns, Andrew W. (Department: 2621)
Image analysis
Image compression or coding
C382S243000, C382S233000, C382S236000, C382S239000, C382S248000, C382S172000, C382S251000, C375S240020, C375S240030, C375S240080, C375S240130, C375S240180
Reexamination Certificate
active
06636637
ABSTRACT:
BACKGROUND OF THE INVENTION
The coding of video signals corresponding to the image coding standards H.261, H.263, MPEG1 and MPEG2 is based on a block-oriented discrete cosine transformation (DCT). In general, these methods use the principle of block-based image coding.
Another approach to image coding is what is called the principle of object-based image coding. In object-based image coding, a segmenting of the image documents takes place corresponding to the objects present in the scene, and a separate coding of these objects takes place.
FIG. 2
shows a general representation of an arrangement for image coding and image decoding.
FIG. 2
shows a camera K with which images are exposed. The camera K can for example be an arbitrary analog camera K that records images of a scene and either digitizes the images in the camera K or also transmits them in analog fashion to a first computer R
1
, in which then either the digitized images B are processed or the analog images are converted into digitized images B and the digitized images B are processed.
The camera K can also be a digital camera K with which digitized images B are recorded directly and are supplied to the first computer R
1
for further processing.
The first computer R
1
can also be fashioned as a separate arrangement with which the method steps specified below can be executed, for example as a separate computer card that is installed in a computer.
The first computer R
1
comprises a processor unit P with which the method steps, specified below, of the image coding or of the image decoding can be executed. The processor unit P is coupled, for example via a bus BU, with a memory SP in which the image data are stored.
In general, the methods specified below can be realized both in software and in hardware, or also partly in software and partly in hardware.
After the image coding has taken place in the first computer R
1
, and after transmission of the compressed image data, via a transmission medium UW, to a second computer R
2
, the image decoding is carried out in the second computer R
2
.
The second computer R
2
can have the same design as the first computer R
1
, i.e. the memory SP that is coupled with the processor unit P via the bus BU.
In
FIG. 3
, a possible arrangement, in the form of a schematic switching diagram for image coding or, respectively, for image decoding, is shown in detailed form, which arrangement can be used in the context of the block-based image coding and partly, as explained below, in the context of the object-based image coding.
In block-based image coding methods, a digitized image B is partitioned into, standardly, square blocks of size 8×8 image points BP or 16×16 image points BP, and is supplied to the arrangement for image coding.
Coding information, e.g. brightness information (luminance values) or color information (chrominance values), is standardly allocated unambiguously to an image point.
In block-based image coding methods, distinctions are made between different image coding modes.
In what is called intra-image coding mode, the overall image is respectively coded with the overall coding information allocated to the image points of the image and is transmitted (I-image).
In what is called inter-image coding mode, only the difference image information of two chronologically successive images is coded and transmitted (P-image, B-image).
Two switching units SE are provided for the changeover between the intra-image coding mode and the inter-image coding mode. For the execution of the inter-image coding mode, a subtraction unit S is provided in which the difference of the image information of two successive images B is formed. The overall image coding is controlled via an image coding control unit ST. The image blocks BB or, respectively, difference image blocks BB to be coded are respectively supplied to a transformation coding unit DCT, in which a transformation coding, for example discrete cosine transformation (DCT), is applied to the coding information allocated to the image points.
In general, however, any transformation coding, e.g. a discrete sine transformation or also a discrete Fourier transformation, can be executed.
The spectral coefficients formed by the transformation coding are quantized in a quantization unit Q and are supplied to an image coding multiplexer (not shown), e.g. for channel coding and/or for entropy coding. In an internal reconstruction loop, the quantized spectral coefficients are inversely quantized in an inverse quantization unit IQ and are subjected to an inverse transformation coding in an inverse transformation coding unit IDCT.
In addition, in the case of inter-image coding, image information of the respective chronologically preceding image is added in an addition unit AE. The images reconstructed in this way are stored in an image memory SP. For simplicity of representation, in the image memory SP a unit for motion compensation MC is shown symbolically.
In addition, a loop filter (LF) is provided that is connected with the memory SP and with the subtraction unit S.
In addition to the image data to be transmitted, a mode flag p is supplied to the image coding multiplexer, which flag indicates whether an intra- or an inter-image coding was executed.
In addition, quantization indices q for the spectral coefficients are supplied to the image coding multiplexer.
A motion vector v is also respectively allocated to an image block and/or to a macro block that contains e.g. 4 image blocks, and is supplied to the image coding multiplexer.
In addition, an information indication f for the activation or, respectively, deactivation of the loop filter LF is provided.
After transmission of the image information via the transmission medium ÜM, the decoding of the transmitted data can take place in the second computer R
2
. For this purpose, in the second computer R
2
an image decoding unit is provided that has for example the design of the reconstruction loop of the arrangement shown in FIG.
2
.
In object-based image coding methods, each image object is first decomposed into blocks of a fixed size, e.g. likewise 8×8 image points. After this decomposition, a part of the resulting image blocks is located completely inside an image object BO. This situation is shown in FIG.
4
. The image B contains at least one image object BO that is outlined with an object edge OK of the image object BO. In addition, image blocks BB with 8×8 image points BP are shown. Image blocks BB that contain at least a part of the object edge OK are called edge image blocks RBB in the following.
Image blocks BB that are located completely inside the image object BO after the decomposition can be coded with a standard block-based discrete cosine transformation, using the above-named block-based image coding method. However, the edge image blocks RBB are partly filled with image information, and must be coded using a separate method.
For the coding of the edge image blocks RBB, up to now there have been two basic approaches.
From a first document, ISO/IEC JTC1/SC29/WG11, MPEG4 Video Verification Model-Version 5.0, Doc. N1469, November 1996, pp. 55-59, it is known to supplement the image information of the image object BO within the edge image block RBB by means of a suitable extrapolation method of the coding information onto the surface of the complete edge image block RBB. This procedure is called padding. The supplemented surface is subsequently coded with a standard 2-dimensional discrete cosine transformation.
Alternatively the first document and a second document, T. Sikora and B. Makai, Shape Adaptive DCT for Generic Coding of Video, IEEE Transactions on Circuits and Systems for Video Technology, Vol. 5, pp. 59-62, February 1995, it is known that the given image object BO is transformed separately according to lines and columns. This procedure is called shape-adapted transformation coding; in the concrete case of the application of a DCT, it is called shape-adapted DCT. The DCT coefficients allocated to the image object BO are determined in such a way tha
Alavi Amir
Johns Andrew W.
Schiff & Hardin & Waite
Siemens Aktiengesellschaft
LandOfFree
Method and arrangement for coding and decoding a digitized... 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 and arrangement for coding and decoding a digitized..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and arrangement for coding and decoding a digitized... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3171073