Determination of Color, Texture and Edge Features for Content-Based Image Retrieval

Expansion of content-based image retrieval (CBIR) system has become a vital research issue as the digital image libraries and multimedia databases are increasing day by day. The image retrieval is divided into two major types: Text-based image retrieval and Content-based image retrieval. Conventional text-based image search engines apply manual annotation of images whereas the content-based image retrieval retrieves images based on image content. This paper explains about the determination of color, texture and edge features required for content-based image retrieval. The image features like color is expressed by using mean and standard deviation, the texture feature is expressed by gray level co-occurrence matrix (GLCM) and the edge features are expressed by using edge histogram descriptor (EHD).