Clustering Algorithms for Image Retrieval from Compressed Images

摘要

Digital images are useful media for storing spatial, spectral and temporal components of information. Large image databases often store the images in compressed format, JPEG for example. This paper examines the algorithms of direct extraction of low level features from compressed images, working with two different clustering techniques. Results indicate that a K-means clustering algorithm has the advantage of a shorter searching time, whereas an Isodata clustering algorithm can have a higher accuracy rate.