Biometric Face Recognition System using SURF Based Approach

摘要: Face recognition can be viewed as the problem of robustly identifying an image a human face, given some database known faces [6]. We propose novel, SURF based approach to face recognition. Although results are not gratifying our proposed loosens burden creating sub spaces is done in PCA, LDA and most recent Bayesian approach. Also, during experiments even though we used unturned program for approach, it outperforms basic PCA approaches terms consistency. This article presents scale-invariant novel rotation detector descriptor (Speeded-Up Robust Features). previously defined schemes with respect repeatability well distinctiveness robustness. It’s computing comparing much faster. This by relying on integral images convolutions; making strengths leading existing detectors descriptors (specifically, using Hessian matrix-based measure detector, distribution-based descriptor); simplifying these methods essential. Its result combination detection, description, finding match steps. The paper contains overview then finds out effects important parameters. concluded SURF’s application two challenging. Yet converse goals i.e. camera calibration which special case registration objects. Our show that very useful vast areas computer vision.