IMAGE PROCESSING AND FACE DETECTION FOR FACE RECOGNITION SYSTEM

• Main Author: (NIM 13203127), ROBIN
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/11320
• Institution: Institut Teknologi Bandung

<p align="justify">Face recognition is a personal identification system which uses a person's personal characteristic (in this case the person's face) to identify the person's identity. The system developed in this final project is a face recognition system which uses FLD (Fisher Linear Discriminant) for feature extraction. The system's designing process is divided into 2 phases: face detection and face recognition phases.<p align="justify"><p>The focus of this paper is on the face detection phase. The input image is a digital image which will be processed through several normalization processes. The normalization processes designed in this project include scale and lighting normalization. The normalization processes are needed to improve the success rate of the classifier. The classifier is used to discriminate between a face image and a non face image. In this final project, SVM (Support Vector Machine) is used as the classifer.<p align="justify"><p>By combining the FLD feature extraction method and SVM as classifier, the system would be able classify part of the input image which contain a face image from the whole input image. The output of the face detection system developed in this project is a face image that is already normalized. This normalized image will then processed by the face recognition system for recognition.<p align="justify"><p>Tests conducted in this final project resulted in 83% (83 correct images detection and recognition out of 100 test images) accuracy rate of the overall recognition system. This accuracy rate is high considering that the system developed in this project is able to identify individuals with different poses from the database.