Host: Japan Society for Fuzzy Theory and Intelligent Informatics
Co-host: International Fuzzy Systems Association, IEEE Computational Intelligence Society Japan Chapter
In this paper, a novel fuzzy-clustering-based-approach for object recognition is proposed. In situations involving multiple objects that can be replaced by a primitive model, the proposed method can be applied without prior information about the number and the shapes of the objects. The approach is composed of three stages. In the first stage, 3D data is reconstructed using stereo matching from a stereo image that includes multiple objects. Next, the 3D data is separated into objects by using a Fuzzy c-Means algorithm augmented with a criterion about the number of clusters. Finally, the shape of each object is extracted by Fuzzy c-Varieties with noise clustering. The effectiveness and validity of the proposed method was shown using both preliminary simulation data and real data obtained from stereo matching.