Sketch based Shopping

Abstract: Content based Image Retrieval is the rising research areas in digital image processing. Sketch Based Image Retrieval is an extension of Content Based Image Retrieval, in which rather than providing a digital image as a query, a hand drawn sketch is used. Sketches are intuitive and descriptive, which offers a more natural way to provide detailed visual clues than pure text. The main objective of this project is to search the similar images from the database where it matches the given input sketch. Database is nothing but large collection of images and those images are extracted and the feature values of these extracted images are saved. The input image is also extracted to get the feature value. This feature value is compared with the values stored in database. If similarity matches, then that image will be displayed as output. Speed up robust feature descriptor is used to make retrieval process fast.

Keywords: CBIR, SBIR, Sobel Operator, SURF.