A New Approach for Clustering of X-ray Images
Chhanda Ray, Krishnendu Sasmal, West Bengal
Medical image databases are a key component in future diagnosis and preventive medicine. Automatic clustering of medical images plays an important role for structuring of a given medical database as well as for searching and retrieval of biomedical images. This paper introduces a new approach for efficient clustering of x-ray images based on various levels of image features. Initially, for each given x-ray image, features have been extracted from three different levels, namely, global, local and pixel. Finally, a new approach, a combination of k-means and hierarchical clustering techniques, has been applied in order to cluster x-ray images. The experimental results are also presented at the end of this paper.