Author : Mr. Majage Vijay and Prof. Vilas S. Gaikwad

Face recognition has increased a noteworthy situ- ation among most usually utilized uses of picture preparing.  With the quick development in media substance, among such substance face acknowledgment has got a lot of consideration particularly in recent years. Face as an item comprises of unmistakable highlights for identification; in this manner,  it  stays most testing research region for researchers in the field     of PC vision and picture handling. Halfway face pictures are created in an unconstrained domain. A face might be blocked   by shades, a cap and a scarf, caught in different postures, situated incompletely out of cameras field of view. Human face assumes a significant job in our social cooperation, passing on individuals’ character yet it is a powerful item and has a high level of inconstancy in its appearances. The issue of perceiving    a self-assertive fix of a face picture remains to a great extent unsolved. This study proposes a new partial face recognition approach, called Dynamic Feature Matching, which combines Fully Convolutional Networks, Principal Component Analysis and Sparse Representation Classification to address partial face recognition problem regardless of various face sizes. DFM does not require prior position information of partial faces against a holistic face.

Keywords: Dynamic feature matching, Partial face recogni- tion, Principle component Analysis, Fully convolutional network.