A GOOGLENET ASSISTED CNN ARCHITECTURE COMBINED WITH FEATURE ATTENTION BLOCKS AND GAUSSIAN DISTRIBUTION FOR VIDEO FACE RECOGNITION AND VERIFICATION

Recently, visual surveillance systems has gained huge attraction from research community due to their significant impact on monitoring application. Several techniques have been developed which are based on the still image, which do not provide efficient solution for real-time application. Hence, video based face recognition is considered a tedious task. Recently, deep learning based schemes have been adopted widely for video face recognition but these techniques suffer from well-known challenges such as pose and illumination variation. Hence, we present a Convolutional Neural network based approach for video face recognition. In this work, we have introduced CNN based scheme which uses feature extraction and feature embedding modules along with GoogleNet architecture to improve the learning of CNN. We have incorporated histogram equalization-based image enhancement technique to improve the quality of video frames. The proposed approach is implemented using Python 3.7. The experimental analysis shows that proposed approach achieves the accuracy as 98.55% and AUC as 99.10% for open source datasets whereas for real-time scenarios