Model for image search and retrieval

Abstract

As a final year project I am presenting an image classification system which is based on the bag of objects principle. The vocabulary is generated using SIFT as the local descriptor and K-means as the clustering algorithm. The main advantage of SIFT is the robustness against noise while with regards to K-means is the simplicity of the algorithm. The classification method used is the Maximum Likelihood which is simple and easy to compute. The main idea of this project is to avoid that daunting and time consuming manual image search in personal computers and devices.