A Novel Framework for the Accuracy Enhancement of Facial Expression Recognition System

Facial Expression Recognition has become a promising field for more natural interactivity with computing devices and machines and has become the focus of attention for many research scholars over the past decade. Newly developed facial emotions recognition methods focus on neutral expression or six expressions used in most state-of-the-art methods. Accuracy is the main problem in the face recognition results. The problem that must be tackled is the optimization of the expression recognition algorithm i.e. to detect, isolate and correctly translate one of the major expressions of the human face with accuracy targeted towards 100%. This work will try to improve the accuracy of recognizing facial expression by using Histogram of Oriented Gradients (HOG) and Local Ternary Pattern (LTP).