Face Recognition Using 3D CNNs

The area of face recognition is one the most widely researched areas in domain computer vision and biometric. This because non-intrusive nature biometric makes it comparatively more suitable for application surveillance at public places such as airports. primitive methods could not give very satisfactory performance. However, with advent machine deep learning their recognition, several major breakthroughs were obtained. use 2D convolution neural networks(2D CNN) crossed human accuracy reached to 99%. Still, robust presence real-world conditions variation resolution, illumination pose a challenge researchers recognition. In this work, we used video input 3D CNN architectures capturing both spatial time information from environment. For purpose experimentation, have developed our own dataset called CVBL dataset. videos shows promising results DenseNets performing best an 97% on