Image analysis – Pattern recognition – Feature extraction
Reexamination Certificate
1999-12-14
2004-01-06
Werner, Brian (Department: 2621)
Image analysis
Pattern recognition
Feature extraction
C382S173000, C382S205000, C382S291000
Reexamination Certificate
active
06674904
ABSTRACT:
BACKGROUND
This invention relates generally to object identification in connection with digital images.
A number of techniques for detecting the edges of objects depicted in digital images are available including the Laplacian, Rule-based, Prewitt and Sobel techniques. Each of these techniques is subject to a number of shortcomings. In many cases, the edge detection algorithms define edges of features that vary too much in proximity to the actual edge of the object. Some of these techniques define thick multi-pixel wide edges. Such techniques make it difficult to find the actual outline of an object because the edges themselves become objects.
Other techniques are prone to developing too many non-continuous edges. Often these algorithms are unable to detect the entire outline or shape of a given object. Instead, they find sections of the overall object's outline.
All of these techniques are susceptible to noise degradation. Noise greatly affects these edge detection algorithms, causing them to produce what are known as false edges. All these techniques are also processor intensive. To obtain greater accuracy translates to more CPU cycles, usually with little improvement. For example, the Rule based algorithm may take up to thirty seconds on one image and the result may still be unacceptable.
Thus, there is a need for better edge detection algorithms and particularly to algorithms which are more amenable to detecting complex shapes and more amenable to detecting shapes in images which are subject to distortion or noise.
SUMMARY
In accordance with one aspect, a method includes receiving digital image information. The contour of an object depicted in the image information is traced. The boundary of said contour traced object is then detected.
Other aspects are set forth in the accompanying detailed description and claims.
REFERENCES:
patent: 4097845 (1978-06-01), Bacus
patent: 5119439 (1992-06-01), Osawa et al.
patent: 5233670 (1993-08-01), Dufour et al.
patent: 5459794 (1995-10-01), Ninomiya et al.
patent: 5579409 (1996-11-01), Vaidyanathan et al.
patent: 5923365 (1999-07-01), Tamir et al.
patent: 5-165963 (1993-07-01), None
Gregory A. Baxes, Digital Image Processing, ISBN 0-471-00949-0, 1994, pp. 98-99.
Intel Corporation
Trop Pruner & Hu P.C.
Werner Brian
LandOfFree
Contour tracing and boundary detection for object... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Contour tracing and boundary detection for object..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Contour tracing and boundary detection for object... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3185606