Face Recognition and Facial Expression Recognition Algorithm Based on Novel Features Extraction Descriptors HDG and HDGG
Farid Ayeche, Adel Alti · 2021
Emotional expressions exchanged through visual communication require a very particular interest particularly to ensure the robust identification of any kind of facial expressions. This problem is becoming harder in unstable environments such as laying, lighting, etc. In this paper, we present novel face and facial expression recognition descriptors called HDG (Histogram of Directional Gradient) and HDGG (Histogram of Directional Gradient Generalized). These descriptors are based on the extraction of local directional gradients. Our main goal is to reach a high face identification accuracy and fast execution time. To reach this goal, we choose to work on two standard databases, JAFFE and YALE. The most recent face descriptors are used to evaluate the efficiency of the presented descriptors. This approach has less complexity since it describe facial expressions by reduced features vectors, which minimizes the computation time. We detail the obtained results by using the well-known metrics to evaluate the recognition rate and execution time of the proposed descriptors.