Indexing and Retrieval of Images Using Wavelets - Segmentation Using Bisecting K-Means Clustering

Using images as identity tool is being implemented in various applications. Images like iris pattern, thumb impressions are being used as identity tools for persons. A sophisticated algorithm is required to index and retrieve appropriate image when queried with a target image. This paper discusses one such image indexing algorithm using wavelets. The images are segmented using Bisecting K-Means algorithm in wavelet domain. It is used to find out appropriate segments in the images and helps for further processing. Feature vectors are extracted and stored in the database for querying and retrieval purpose. Query image is also processed in the same method and the feature vector is extracted. This vector is compared against vectors stored in the database and relevant images are retrieved based on the similarity value.