Due to the explosion of multimedia quantities such as archived TV broad cast videos, various multimedia resources released on the Internet, intense work in multimedia retrieval domain has aimed to provide efficient and accurate functionality for users to access the desired information. The rapid access to these huge multimedia resources and automatically assess day similarity or an experience from this collection of unstructured egocentric data presents major challenges that require efficient algorithms. As everything is nowadays automated, the main goal of this work is to apply a new method to automatically assess day similarity using a real image dataset called Egocentric Dataset of the University of Barcelona (EDUB) of 4912 daily images acquired by 4 persons in order to evaluate the performance of our system.