DESIGN OF FACE RECOGNITION SYSTEM FOR CANDIDATE VERIFICATION PROCESS PRIOR TO THE EXAMINATION

Geevitha T, Hemanth Kumar K S, Siva Yellampalli

Abstract: A face recognition system which is designed in this project is mainly based on facial geometry measurement. The design work can be divided as two sub-stages, face detection and face recognition. The face detection part of the project will be achieved by the viola jones detector and face recognition part of the design will be done by using artificial neural networks. The developed system can also detect and performs the recognition of multiple faces which are captured in free environment and also recognizes the face from the stored test database. This project is an attempt to work on the problem of human face recognition during candidate verification process prior to the examination. The automated face recognition is implemented using the Matlab technical computing language. The frontal view facial geometry based face recognition system is also expanded and implemented and tested on facial images for different subjects with different poses. The face database of 10 individuals consisting of each 10 different facial images, hence 100 facial images in database to test automated face detection. This project represents the development of a system which can identify the person with the help of a facial features using artificial neural network technique. The features in the image are identified and distance between such identified features are measured. Such measurements are represented as a matrix and this feature matrix is given as input and fed to artificial neural network. The designed system for data base as input could attain 100% rate of recognition accuracy where as for real time input, rate of recognition accuracy is 96%.

Keywords: Facial Geometry Measurement, Face Detection and Face Recognition, Viola Jones Detector, Artificial Neural NetworkAX

DOI: https://doi.org/10.15623/ijret.2016.0516078