Recognizing Face Using the Combination of Singular Value Decomposition and Hidden Markov Model Algorithms

Face recognition as a biometric system is a development of an authentication system based on the face’s natural characteristics. Face recognition needs to be further researched because it has been applied to various fields. After all, this system does not require direct physical contact between humans and computer input sensors. However, problems in face recognition systems are pretty dynamic and complex. This study proposed a face recognition system using Singular Value Decomposition (SVD) for feature extraction and Hidden Markov Model (HMM) for classification. It is expected that this research may contribute to developing the method of face recognition. The recognition process is based on a frontal face image which is divided into seven areas, and each area is assigned to a state in a one-dimensional HMM. 100 images are used for database creation. Image acquisition is made 10 times for each of 10 different people with different expressions. Half of the images are used for training and the rest for the testing process. In this study, the moving probability from one state to the state itself and the next state is not the same. Still, it depends on the average value of the height of the area covered by a particular state, and the result shows that it can increase the recognition rate from 94 to 96%.