An effectual face tracking based on transformed algorithm using composite mask
G. Ramkumar, E. Logashanmugam · 2016
The cameras which use in day to day life are low resolution and high resolution cameras. In Practical video scrutiny applications, the faces developed by outside cameras are of low degree resolution and are disturbed by uncontrolled brilliance. Since low resolution cameras are preferred for surveillance because of its low cost. The existing GLF feature is not robust to track face at large changes in lighting direction. In order to overcome this issue, a MGLF (Modified Gradient Logarithm Field) which chiefly depends on the intrinsic appearances of a face and only slightly affected by the lighting source is proposed with the help of novel filter. Also this method does not stand on a particular face model. Output shows that implemented using a composite mask, the MGLF tracker is not sensitive to illumination variations and outstrips many state-of the art pursuing algorithms.