Facial Expression Recognition via Modified GAD Features with PSO-KNN
I. Michael Revina, W. R. Sam Emmanuel · 2018 International Conference on Smart Systems and Inventive Technology (ICSSIT) · 2018
In Image processing, Facial Expression Recognition (FER) is the exigent and remarkable effort as well as realized through computers otherwise human being. This paper suggests the Local Descriptor (LD) with Modified Gray value Accumulating Distance (MGAD) along with Particle Swarm Optimization based K-Nearest Neighbor (PSO-KNN) classifier for efficient FER. This FER system contains two major stages are LD-MGAD feature extraction and PSO-KNN classification of facial expressions. In the first stage, the natural and powerful metric methodology of MGAD is calculated. Subsequently, based on MGAD the LD is formed that vigorously gains the local structure knowledge among the middle and its adjoining pixels. In the second stage, the LD-MGAD features are optimized through the PSO algorithm. The selected feature vectors are classified using the KNN classifier. The modified GAD features with PSO-KNN inspired FER system to identify the facial expressions accurately that are happy, sad, surprise, disgust fear, and anger. The experimental results of the proposed method are evaluated using the facial expression database JAFFE also it attains better recognition accuracy rate.