Hybrid Intelligent System for Face Recognition from Surveillance Camera

Intelligent systems have been widely used for face recognition. Among them Support Vector Machine (SVM) was recognized as a powerful recognition model. However, handling the problem of recognizing a faces from surveillance camera is difficult task due to that it encounter a variation in pose, resolution, as well as illumination. In this work, evolutionary constructed SVM-based intelligent system will be developed. Particularly, the developed system comprises the hybridization of Gray Wolfe Optimizer (GWO) [1] with SVM. DE is used to construct an efficient SVM recognition model by performing simulations parameters tuning, training instances selection, and features selection. To evaluate the performances of the presented model, a number of benchmarks for surveillance-based face recognition problem will be used such as ChokePoint, UCSD/Honda, CMU, and YouTube Faces (YTF) database.