Image analysis – Pattern recognition
Patent
1997-10-14
2000-07-18
Johns, Andrew W.
Image analysis
Pattern recognition
G06K 962
Patent
active
060918527
DESCRIPTION:
BRIEF SUMMARY
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a technique for pattern recognition, in which the pattern of an n dimensional object is measured and recognized by means of a data-processing system capable of mapping and processing the n-dimensional object with different accuracy (s) n computer memory. Further subject matter of the invention is a technique for generating or modifying an n-dimensional object being simulated or represented by a data-processing system, the data-processing system being capable of simulating or representing the n-dimensional object with different accuracy (s).
2. Description of the Background
Pattern recognition is a central problem in many technical fields. Such pattern recognition is supposed to allow n dimensional objects to be sufficiently measured by means of a data-processing system so that the data-processing system can identify with the highest possible precision to which n-dimensional object the respective pattern must be assigned. Highly precise pattern recognition of n- dimensional objects would make it possible, for example, to steer vehicles of any kind automatically, so that accidents due to human error could be largely prevented. It would also be possible to recognize the handwriting of any given person automatically and with high precision. Even the manufacture of automatic machines or robots equipped with intelligent sensor systems would be no problem if highly precise pattern recognition were available.
Numerous pattern-recognition techniques are already known from the prior art. One disadvantage of the known techniques, however, is that each is usable only for special types of objects. The general usability of these known techniques is therefore greatly restricted. A further disadvantage of the known techniques, particularly for highly complex or fractal structures, is that reliable recognition is not possible at all or requires extremely large computing power, and so real-time pattern recognition is not possible.
The problem of simulating and/or representing n-dimensional objects by means of a data-processing system is now being encountered in numerous technical fields. For example, in the object-oriented drawing, image-processing, vector-graphics and CAD programs now in vogue, it is necessary to modify certain selected objects appropriately by selective manipulation. The techniques known heretofore for selective modification or initial generation of such an object are often user-unfriendly and laborious. In particular, rapid generation or modification of objects having highly complex structure is difficult for these known techniques.
The objective of the invention is to improve a pattern-recognition technique in which the pattern of an n-dimensional object is measured by means of a data-processing system capable of mapping and processing the n-dimensional object with different accuracy(s) in computer memory, and an object-manipulation technique for generating or modifying an n-dimensional object being simulated or represented by a data-processing system, the data-processing system being capable of simulating or representing the n-dimensional object with different accuracy (s), to the extent that a universally usable technique for rapid and precise pattern recognition and for extremely simple manipulation of the n-dimensional object in question is ensured.
This objective is achieved according to the invention with the steps of the pattern-recognition technique characterized in that: in computer memory, the n-dimensional object is determined by integration or summation, and in that characterizing variable(s) on accuracy(s) is determined and used as a measure of the pattern to be recognized. (U) on accuracy (s) is input beforehand, the characterizing variable (U) corresponding to a property of the n-dimensional object obtainable by integration or summation, and in that input functional dependence (U(s)).
SUMMARY OF THE INVENTION
The basic principle of the invention therefore comprises identifying an n-dimensional objec
REFERENCES:
patent: 5793371 (1998-08-01), Deering
patent: 5809183 (1998-09-01), Serizawa et al.
Leu, "Computing a Shape's Moments From Its Boundary," Pattern Recognition, vol. 24, No. 10, 1991, pp. 949-957.
Entleitner Renate
Soegtrop Michael
Delphi 2 Creative Technologies GmbH
Johns Andrew W.
LandOfFree
Method of pattern recognition and method of creating an n-dimens 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 pattern recognition and method of creating an n-dimens, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of pattern recognition and method of creating an n-dimens will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2045686