Image Retrieval by Content Using Descriptors

Image retrieval from large databases has become easier nowadays through the digitization of images. One such methodology which is widely used forms the image retrieval by its content. CBIR enhances the technique of retrieving images from a larger collection. QBIC has achieved its necessity by retrieving the appropriate images by using its features to retrieve the images. Features can be of one of the following- Shape, Color and Texture. This paper makes use of descriptors to retrieve the texture and the color features. Texture feature is extracted by making use of Edge histogram and color feature is extracted by using HMMD color space. The feature extraction phase is applicable for both query image and database images. These features are processed and stored as another database named feature database which is later used to match the query image with the feature database images. Similarity measure is calculated between the images through a distance measure and hence similar images are retrieved that matches the specific criteria.