Research On Face Detection Algorithm Based On Deep Learning

Face detection technology is one of the hot topics in the field of computer vision.The technology is to use information means to locate the face in the media such as pictures or videos,and mark its position and size through tags.The development of this technology is the basis of face recognition algorithm research,directly determines the final recognition effect of face recognition algorithm.However,in the actual face detection process,it is often affected by the complex detection scene,face occlusion,low resolution,face deflection and other human and environmental factors,resulting in the application of traditional detection methods in the real environment can not achieve accurate,fast,real-time effect.In recent years,with the development of artificial intelligence technology,especially the application of deep learning technology to face detection technology in the detection speed and accuracy of revolutionary changes,and has been greatly improved.In this paper,on the basis of deep learning theory,face detection algorithm is further studied.By applying deep learning algorithm theory to face detection algorithm,normal face and abnormal behavior are detected respectively.The main contents are as follows:1.This paper reviews the traditional face detection algorithms convolutional neural networks and deep learning based face detection algorithms in detail.Two representative algorithms MTCNN and S3 FD are introduced.2.In view of the shortcomings of fast convolutional neural network algorithm in processing small face detection,such as low detection accuracy and low speed,this paper optimized and improved the existing algorithm based on fast convolutional neural network and proposed an improved algorithm based on fast convolutional neural network for face detection.Finally,the proposed algorithm has been validated on Wider FACE and FDDB datasets,and it is proved that the algorithm achieves good results in small FACE detection.3.The fast convolutional neural network is further studied and applied to video abnormal behavior detection,and an abnormal behavior detection algorithm is proposed.4.The research work of this paper is summarized and prospected.