Content-Based Image Retrieval using color and shape descriptors

Content-Based Image Retrieval technique uses three primitive features like color, texture and shape which play a vital role in image retrieval. This paper presents a novel framework using color and shape features by extracting the different components of an image using the Lab and HSV color spaces to retrieve the edge features. Invariant moments are then used to recognize the image. In this proposed work, the performance of the HSV and Lab color space approach have been compared with Gray and RGB approach. Accordingly the Lab color space approach gives better performance than RGB and HSV. The experiments carried out on the bench marked Wang's dataset, comprising Corel images, demonstrate the efficacy of this method.