Survey of content-based image retrieval using SIFT algorithm- case study (image/sketch retrieval)

Abstract

In Digital Image Processing area, Content Based Image Retrieval is growing in popularity. Google and Yahoo have tools on Digital Image Processing. Those tools are based on textual image annotation. In textual annotations in addition to the key-words images are obtained. This method is quite ineffective as its performance isn’t fulfilling. The content based Image retrieval depends on automatic extracting of content according to color, texture, and so on. This paper is focused on discussing obstacles in the improvement of Content Based Image Retrieval based on free hand sketch (“Sketch Based Image Retrieval”). The study will be focused on trying creating task specific descriptor for handling the information gap existing between colored images and sketches and that will give the chance for efficient search. The descriptor is formed after such special series of preprocessing stages which the sketch and converted images may be compared. SIFT is the topic covered. The study will explain the way SIFT operates and its advantages in addition to the fact that SIFT is more efficient in the area of image retrieval.