A Review on Unconstrained Real-Time Rotation-Invariant Face Detection

Abstract

With the amazing growth of image and video databases, there is a vast need for intelligent systems to automatically understand and look at information since doing it by hand is getting very hard. Faces are significant in social interactions because they show the feelings and identity of a person. People are not much better than machines at recognizing different faces. The automatic face detection system is a key in head pose tracking, face verification, face recognition, face tracking, face animation, face modeling, facial expression recognition, age and gender recognition, and behavior analysis in a crowd. Face detection is a way for a computer to find out the size and location of a face in an image. Face detection has been an outstanding issue in computer vision literature. This paper provides an overview of pose and rotation invariant face detection approaches with architecture designs and performance on popular benchmark datasets. The benchmark datasets used for face detection are listed as their key features. This paper also talks about different applications and challenges with face detection. Also, we set up special discussions on the practical aspects of making a face-detection system that works well. We end this paper by suggesting a few promising directions for future research.