Individual Recognition System Through Face Identification Using Hidden Markov Model
7 Pages Posted: 25 Sep 2018
Date Written: August 1, 2018
Biometrics is a branch of applied mathematics that is used for the purposes of individual identification through fingerprint, retina eye, palm, voice or face. The example of biometrics application is the identification through individual faces. In this study, we use Hidden Markov Model (HMM) for face identification. The process consists of image retrieval, image preprocessing, block extraction, feature extraction, training process, and identification process. The testing images consist of 120 face images that are divided into two types: 80 images collected from non-moving objects and 40 images collected from moving objects (video). The highest accuracy is obtained from image of size 64 × 64 pixels with T = 59. The accuracy of the system w.r.t. training images is 98.13%, and the accuracy of the system with all testing images is 80.83%, consisting of 85.00% for the accuracy of the non-moving face and 72.50% for the moving face.
Suggested Citation: Suggested Citation