Topology representing networks
Neural Networks
Self-Organizing Maps
IEEE Transactions on Pattern Analysis and Machine Intelligence
An empirical approach to grouping and segmentation
An empirical approach to grouping and segmentation
Toward Objective Evaluation of Image Segmentation Algorithms
IEEE Transactions on Pattern Analysis and Machine Intelligence
Image segmentation evaluation: A survey of unsupervised methods
Computer Vision and Image Understanding
A Color Based Image Segmentation and its Application to Text Segmentation
ICVGIP '08 Proceedings of the 2008 Sixth Indian Conference on Computer Vision, Graphics & Image Processing
RSFDGrC'03 Proceedings of the 9th international conference on Rough sets, fuzzy sets, data mining, and granular computing
Color Image Segmentation Based on Mean Shift and Normalized Cuts
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
CTex—An Adaptive Unsupervised Segmentation Algorithm Based on Color-Texture Coherence
IEEE Transactions on Image Processing
Clustering of the self-organizing map
IEEE Transactions on Neural Networks
| Hi-index | 0.00 |
Topology adaptive neural networks are popular because of its capability to adopt the underlaying topology of data. In this paper we develop a topology adaptive self organizing circular neural network (TASOCNN) model for circular-linear data sampled from an unit disk. The basic framework uses the TASONN procedure of Datta et. al. [8]. The update rules and the distance measure are reformulated with the inclusion of directional information. In the iterative TASOCNN process, we create/update edges between any two winner processors. These edges are weighted, hence finally a weighted processor graph is created concerning processors as vertices. By removing possible inter-cluster edges from the connected components of this processor graph, a set of subgraphs can be obtained. For this purpose we use a cost function based on edge length and strength. Many of these subgraphs become close to one another. These close subgraphs are merged. Finally, each subgraph represents a cluster in the data. We apply TASOCNN for color based image segmentation. TASOCNN is constructed in the hue-saturation space. The Berkeley segmentation dataset is used to present the results. An evaluation is made by means of probabilistic rand index. Our experiments reveals satisfactory outcome of TASOCNN.