Abstract:

Abstract The current emphasis of computer vision research is face recognition. This system has one of the fastest recent growth rates among biometric systems. Several studies have been conducted to identify facial photos using a variety of methodologies, including appearance-dependent methods, feature-based methods, and hybrid methods, with varying degrees of success. This research provided a method for classifying and extracting features predicated upon deep learning for face recognition. The data was first prepared for noise reduction and picture resizing, after which the image was segmented for smoothing. After that, employing Scale Invariant Feature Belief Network Transforms, extract the features with classification. Given its high accuracy, it appears that deep learning a viable method for doing facial recognition. The classified output shows face features and The results of a parametric analysis have been done with regards of accuracy, precision, recall and F-1 score for face dataset. Suggested SIFBNT accomplished accuracy of 95%, precision of 76.5%, recall of 86%, F-1 score of 79%

Keywords: