Abstract

- Content based image retrieval upgrading more rapidly with increasing of Electronics, Informatics and Telecommunications equipments, it can overcome the drawback of image retrieval based on text symbols. - In this trend, we represent one model to retrieve image based on content. It can be used by the committee having own image and video collections, digital libraries, satellite video, medical imaging, education and distance learning. - Methodology based on using image features as color, textures, shape, spatial relationships, to classifying image databse, attaching semantic tags to image for efficiently retrieving. Retrieving is really changing to classiffication. - The main purpose of this paper is to represent the automatically low-level features extracting, using Bayesian classification based on vector quantization to estimate unknown probability density function from the training data. - Extracting high-level conceptual information from low-level images features is a challenging problem. Automatically extracting all such relevant information may not be possible. Hower certain semantic concepts can be identified in a constrained environment.