Pose invariant face recognition using HMM and SVM using PCA for dimensionality reduction

An embedded system is presented in which face recognition and facial recognition for Human-Robot Interaction are implemented. To detect face with a fast and reliable way, HMM combined with SVM algorithm is used. The full pose from face recognition data base is considered to detect the face recognition. Performance of the face recognition reaches to 99.617%. The two main advantages of our method are that it does not require manually selected facial landmarks or head pose estimation. In order to improve the performance of our pose normalization method in face recognition, an algorithm is presented for classifying whether a given face image is at a frontal or non frontal pose. Experimental results on different datasets are presented to demonstrate the effectiveness of the proposed approach. In addition to the proposed method, a pre processing state of detecting the face is included here. The images which are non faces are detected and eliminated from the database is an another main advantage in of this work.