Recent trends in hierarchic document clustering: a critical review
Information Processing and Management: an International Journal
Parallel fuzzy c-means cluster analysis
VECPAR'06 Proceedings of the 7th international conference on High performance computing for computational science
Constructive feedforward ART clustering networks. I
IEEE Transactions on Neural Networks
Constructive feedforward ART clustering networks. II
IEEE Transactions on Neural Networks
Characterization of Ag-PEO10LiCF3SO3-polypyrrol-Au neural switch
WSEAS Transactions on Computers
Hi-index | 0.00 |
Optical-flow field analysis is one of the most efficient tools for segmenting moving objects in an image sequence, especially when a camera itself is also moving. Object segmentation from the optical flow can be considered as a clustering problem. The performance of clustering method can significantly improve results of moving object segmentation. This work presents the unsupervised clustering system using a modified self-organizing feature map (MSOFM) neural network. The network can automatically perform clustering without having any priori knowledge of any initial number of clusters or any initial spatial position. It also can be adjustable to achieve multi-resolution clustering. This allows the proposed network to segment flows of multiple moving objects having nearly same speeds. The system shows desirable results of segmentation of moving objects in the camera-moving image sequence. Results and discussions of adjustable capability of the network are also presented.