UP3: User profiling from Profile Picture in Multi-Social Networking

Profiling Online Social Network (OSN) Users by matching their Profile Pictures in Multi-Social Networking requires their own frontal face images in consideration. Present State-of-the-Art algorithms are ineffective in detecting mouth and nose on the face, making it inefficient to be used in matching different faces by localizing their facial features. This work proposes a novel approach to improve the effectiveness and efficiency of face detection by bifurcating the detected face horizontally and vertically. The algorithm runs only on the portion of the detected face Bounded Box (BB) to generate bounded boxes of other facial objects, and later the Euclidian distance between the BBs with respect to that of the face is computed to get Logarithm of Determinant of Euclidian Distance Matrix (LDEDM) in Relative-Distance method and stored in the database. The LDEDM so computed is unique for the user image under consideration and is used for the purpose of matching the identity of the user images from the database. The Equal Error Rate (EER) is considerably low with the proposed User Profiling from Profile Picture (UP3) algorithm indicating better performance.