Facial Expression Recognition in ATM Surveillance System

Facial expression analysis is a novel and challenging field with significant applications in data-driven animation and human–computer interaction, among other fields. Facial emotion identification requires a vital step called efficient facial representation, which is created from the original face images. In this paper, Face Expression Recognition (FER) is determined to understand the situation of a human in an ATM environment that may lead to suspicious events in the worst case. A realistic evaluation of Local Binary Patterns (LBP), a facial representation method based on statistical local features. A number of machine learning techniques were investigated in detail on various databases. To extract histograms and Local Binary Patterns (LBP), the face area is first separated into small areas. This feature vector makes a face easily recognizable and helps to find similarities between images. Support Vector Machine (SVM) was used to configure the FER system for multi-class facial expression categorization mode.