Efficient Retrieval of Face Image from Large Scale Database Using Sparse Coding and Reranking

Due to largely mounting of photo sharing in social network services, there is a strong need for large scale content-based face image retrieval which is enabling device for many emerging applications. It is very exigent to find a human face image from large scale database which might hold huge amount of face images of people. Existing methods regulate the location and illumination differences between the faces and abolish background contents while retrieving face images, which leads to decline the important context information. This paper intends to utilize automatically discovered human attributes that contain semantic signs of the face photos to recover large scale content-based face image retrieval. In this paper two methods are used to develop image retrieval in both offline and online stages, they are attribute-enhanced sparse coding (ASC) and attribute-embedded inverted indexing (AEI). Reranking is further used with these two methods to attain significant retrieval outcome. Proposed system exploits automatically discovered human attributes which balance the information loss and attain good retrieval performance compared to existing system.