Texture Based Image Classification and Retrieval Using Fuzzy Clustering and HMM Classifier Approach

Image classification is major area of research. Many of the researchers doing research in this area to find optimal methods or algorithms to classify the query image into relevant image or mimages. The proposed work focuses on textures based image classification and retrieval. Initially, the image features are extracted from images and reduced dimensions of image features using principal component analysis (PCA) method. Then the segmentation process involves grouping of the similar texture components into several groups. We used Fuzzy C-Mean clustering (FCM) to segment the image. So, we get the required sub-image feature for classification. During the classification, Hidden markov model (HMM) is used to match the unknown texture against different set of mimage classes. Finally the best match is taken as the classification result.