Image Clustering by Neural Network (SOM) using Contents of Color and Texture

The rapid development in computer technology for multimedia databases, digital media results in increase in the usage of digital images. Vast amount of data can be hidden in the form of digitized image. Image mining is used to extract such kind of data and potential information from general collections of images. Image Clustering groups the images into classes of similar images without prior knowledge. CBIR has extensive potential applications. Visual content of still images are used by CBIR to search for similar images in large scale. Thus the search for the relevant information in the large space of image and video databases become more challenging and interesting too. This paper discuss the concept of image clustering by self-organizing map (SOM) using the contents color and texture as image features for improving user interaction with image retrieval systems . The visual content of an image is analyzed in terms of low-level features extracted from the image. For color feature extraction, HSV color model and texture Gabor filter is presented.