Abstract | ||
---|---|---|
A framework for the detection of edges on color edges, which is based on the application of the fuzzy integral, is proposed herein. The framework makes use of the confidence measure of [1] in order to automate the construction of the fuzzy measure coefficients. The computation of the confidence measure is achieved by applying a competitive learning algorithm on the input images, whereby the adaptability of the mentioned algorithm is increased. This is not the only advance attained herein, since the automation of the fuzzy measure's construction furthers the application of the fuzzy integral in computer vision. The framework has been applied in the feasibility study of a system for the reconstruction of frescos. The results in this industrial application are shown together with the performance evaluation on some benchmark images. |
Year | DOI | Venue |
---|---|---|
2006 | 10.1109/FUZZY.2006.1681848 | 2006 IEEE INTERNATIONAL CONFERENCE ON FUZZY SYSTEMS, VOLS 1-5 |
Keywords | Field | DocType |
competitive learning,edge detection,feasibility study,unsupervised learning,fuzzy set theory,computer vision | Adaptability,Pattern recognition,Edge detection,Fuzzy set operations,Computer science,Fuzzy logic,Automation,Fuzzy set,Unsupervised learning,Artificial intelligence,Machine learning,Computation | Conference |
ISSN | Citations | PageRank |
1098-7584 | 0 | 0.34 |
References | Authors | |
17 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Aureli Soria-Frisch | 1 | 83 | 11.13 |
Abderrahim Kassid | 2 | 0 | 0.34 |