Automatic face recognition and analysis using soft computing methods

• Main Author: Sim, Fuh Cherng
• Other Authors: Wang Han
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/78258
• Institution: Nanyang Technological University

Automatic facial recognition is the uprising biometric system that is increasingly being implemented for use in today’s world. Many of these biometric systems are being widely used for authentication and identification purposes. Not only that it can be used for authentication and identification purposes, correctly identifying and classifying of facial expressions using automatic facial recognition system can also play an important role in the national security. With the increasing security threats around the world, correctly and quickly identifying a person who is of a target of interest is crucial to the national security. By analysing the mood of a person using their facial expressions, it can act as an early warning indication that the person may be a possible threat in terrorism. This may allow the relevant authorities to conduct a timely investigation on the suspected individual. Therefore, correctly classifying and identifying the mood of a person using their facial expression can also plays an important role in today’s context for counter-terrorism. An important way of communication between human beings is also by facial expressions. The emotion can be understood by another just by looking at their facial expression. Although human can effortlessly recognize facial expressions quickly and accurately, the reliance on machines to correctly identify the type of emotion or expression in a timely fashion is still a great challenge. Automatic facial expression recognition can be important in human machine interface studies as well as in psychological studies such as in human behavioral science. There are currently several known methods used for facial expression recognition such as the eigenfaces method, Principal Component Analysis (PCA) method, Independent Component Analysis (ICA) method, Neural Networks and many more. The recognition of the emotions is done by deploying using MATLAB’s neural network training toolbox. The network classifier is using of pattern recognition neural network which is a feed forward neural network that trains the images bearing different emotions. The trained network is then simulated using live video acquisition of the webcam to test the new data for recognizing the different expressions.