Image analysis – Image compression or coding – Interframe coding
Reexamination Certificate
1998-10-20
2001-08-07
Couso, Jose L. (Department: 2724)
Image analysis
Image compression or coding
Interframe coding
C382S238000, C382S232000, C382S242000
Reexamination Certificate
active
06272254
ABSTRACT:
BACKGROUND OF THE INVENTION
New moving image compression methods are based on a resolution of the image content into what are referred to as image objects with arbitrary edging. The individual image objects are separately encoded in various video object planes (what are referred to as VOPs), and are transmitted and decoded in the receiver and again “combined”. Traditional image compression methods are based on a resolution of the entire image into quadratic image blocks. This principle can also be accepted in the object-based methods. However, problems arise in the encoding of the image blocks that are located at the edge of the respective image object since the object edging usually does not coincide with the block edges. What is referred to as a motion-compensated prediction of these edge blocks is especially critical in this context.
There are a number of different search strategies for the motion estimation and the motion compensation. What is referred to as the “block matching method” is usually utilized for block-based image compression methods. It is based thereon that the image block to be encoded is compared with same-sized blocks of a reference image. One of the reference image blocks is located at the same position as the image block to be encoded; the other reference image blocks are topically shifted compared to it. Given a large search area in the horizontal and vertical directions, a great number of search positions thus derive, so that correspondingly, many block comparison (“matchings”) must also be implemented given what is referred to as a complete search (“full search”). The sum of the absolute differences of the encoding information that is respectively allocated to each picture element is usually employed as a criterion for the match quality between the block to be respectively encoded and the reference block.
What is to be understood by encoding information in the framework of this document is, for example, the luminance information with which a luminance value that a picture element exhibits is respectively described or, too, a color information, i.e. the chrominance value allocated to the respective picture element.
It is known in the motion compensation from the publication ISO/IEC JTC1/SC29/WG11, Coding of Moving Pictures and Associated Audion Information, MPEG 96/N1380. MPEG5 Video Verification Model, Version 4.0, pages 31 through 35, Chicago, October 1996, to implement what is referred to as a padding in the framework of the motion compensation, whereby the padding given the method from the ISO publication occurs over the entire area of the reference image RB that does not belong to the image object BO.
The block-based image encoding method MPEG 2 is known from the publication J. De Lameillieure and R. Schäfer, MPEG-2-Bildcodierung für das digitale Fernsehen, Fernsehund Kino-Technik, Volume 48, No. 3, pp. 99-107, 1994. The publication Xiaolin Wu and Yonggang Fang, A Segmentation-Based Predictive Multiresolution Image Coder, IEEE Transactions on Image Processing, Vol. 4, No. 1, pp. 34-46, January 1995, discloses a method for segmentation of image objects from an image.
SUMMARY OF THE INVENTION
The invention is based on the problem of specifying a method for the image encoding, a method for the image decoding, as well as an arrangement for the implementation of the methods with which the image encoding or the image decoding is possible with a computing time requirement that is reduced compared to known methods.
The problem is solved by the method for image encoding of a digitalized image having an arbitrary number of picture elements wherein the image is provided with at least one image object that comprises an arbitrary number of picture elements. The image is also provided with image blocks. A motion compensation is implemented for the image blocks. A path is implemented in the motion compensation of an object edge image block that comprises at least one object edge of the image object. The padding is only implemented with respect to picture elements that are located within the image object.
Also, according to a method of the invention for an image encoding of a digitalized image having an arbitrary number of picture elements, at least one image object that comprises an arbitrary number of picture elements is identified in the image. The image is divided into image blocks. A motion estimation is implemented for the image blocks. A padding is implemented in the motion estimation of an object edge image block that comprises at least one object edge of the image block. The padding is only implemented with respect to picture elements that are located within the image object.
Also according to the invention, an apparatus is provided for image encoding of a digitalized image having an arbitrary number of picture elements. A processor unit is configured such that: at least one image object that comprises an arbitrary number of picture elements is identified in the image; the image is divided into image blocks; a motion estimation is implemented for the image blocks; a padding is implemented in a motion estimation of an object edge image block that comprises at least one object edge of the image block; and the padding is only implemented with respect to picture elements that are located within the image object.
Also according to the invention, an apparatus is provided for image decoding of a digitalized image having an arbitrary number of picture elements wherein a processor unit is configured such that: the image comprises at least one image object that comprises an arbitrary number of picture elements; the image comprises image blocks; a motion compensation is implemented for the image block; a padding is implemented in the motion compensation of an object edge image block that has at least one object edge of the image object; and the padding is only implemented with respect to picture elements that are located within the image object.
At least one image object is determined in the image given the method for image encoding. Further, the image is demarcated into image blocks and a motion estimation is implemented for the image blocks. A padding is implemented in the motion estimation of an image block that comprises at least one object edge of the image object, whereby the padding is merely implemented with respect to picture elements that are located within the image object.
In the method for image decoding, a padding is likewise implemented only with respect to picture elements that are located within the image object for motion compensation of an image block that exhibits at least one object edge of an image object.
In the arrangement for the implementation of at least one of the methods, first, an image store for storing the digital image data is provided, as is a processor unit for the implementation of the individual methods steps of either the method for image encoding or, too, the method for image decoding.
Both methods as well as the arrangement exhibit the considerable advantage that a considerable saving in computing time is achieved in the implementation of the image encoding or, of the image decoding. This is to be attributed thereto that the picture element supplementation required for the motion estimating or, for the motion compensation is no longer carried out independently of a contour information at a fixed block grid but as mush image information is respectively supplemented as required taking the momentary motion vector and the known object contour into consideration.
It is advantageous in the method for image encoding to implement the padding in the framework of the motion compensation for the internal reconstruction of the image in the same, above-described way.
A further saving of computing time in the image encoding is achieved with this procedure since the advantageous procedure is also implemented in the internal reconstruction.
Further, it is advantageous in the method for image encoding to store values of coding information of previously padded picture elements and to re-employ further image blocks in the motion estimating. A
Couso Jose L.
Do Anh Hong
Schiff & Hardin & Waite
Siemens Aktiengesellschaft
LandOfFree
Method for encoding and decoding of a digitalized image 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 for encoding and decoding of a digitalized image and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for encoding and decoding of a digitalized image and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2437627