Analysis of Multi-View Face Detection using Deep Convolutional Neural Networks

Detecting multi-view faces in any image with a complex and noisy background remain a challenging problem. In last few decades, deep learning especially, convolutional neural networks provided significant improved performance in computer vision. In this paper, we have developed deep convolutional neural networks framework for detecting multi-view faces. Implementation, detection and retrieval of faces is obtained with the help of direct visual matching technology. Further, the probabilistic measure of the similarity of the face images is done using Bayesian analysis. Experiment detects faces with ±90 degree out of plane rotations. Fine tunedAlexNet is used to detect multi view faces. For this work, we extracted examples of training from AFLW (Annotated Facial Landmarks in the Wild) dataset that involve 21K images with 24K annotations of the face. Finally, we show that deep learning architecture of CNN with different structure produces significantly better results on large pose variant face.