Abstract | ||
---|---|---|
From the objective function of a generalized entropy-based fuzzy c-means (FCM) clustering, an algorithm was derived, which is a counterpart of Gaussian mixture models clustering. A drawback of the iterative clustering method is the slow convergence of the algorithm. Miyamoto et al. derived a hard clustering algorithm by defuzzifying the FCM clustering in which covariance matrices were introduced as decision variables. Taking into account this method, for quick and stable convergence of FCM type clustering, we propose the semi-hard clustering approach. The clustering result is used for a classifier and the free parameters of the membership function of fuzzy clusters are selected by particle swarm optimization (PSO). A high classification performance is achieved on a vehicle detection problem for outdoor parking lots. |
Year | Venue | DocType |
---|---|---|
2010 | INTEGRATED UNCERTAINTY MANAGEMENT AND APPLICATIONS | Conference |
Volume | ISSN | Citations |
68 | 1867-5662 | 0 |
PageRank | References | Authors |
0.34 | 0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hidetomo Ichihashi | 1 | 370 | 72.85 |
Akira Notsu | 2 | 146 | 42.93 |
Katsuhiro Honda | 3 | 289 | 63.11 |