The Detection And Location Of The Standing Face In The Intelligent Recording And Broadcasting System Based On GoogLeNet

With the continuous popularization of education informationization,the development of intelligent recording and broadcasting systems has been promoted.The intelligent recording and broadcasting system can promote the construction of high-quality courses in schools,greatly promote the depth of school teaching research and management research,accumulate rich digital teaching resources for the further development of education informatization,and realize sharing of teaching resources.In the teaching process,in order to follow the teacher-student interaction in real time,it is necessary to detect and locate standing students.Therefore,the face detection and positioning of standing students is the key to developing an intelligent recording and broadcasting system.Aiming at the needs of students’ standing detection in the current intelligent recording and broadcasting system,this paper implements a method for detecting and locating standing faces based on GoogLe Net neural network.The method first uses the frame difference method and skin color detection to initially determine the student’s standing activity area,and then uses the transfer learning training GoogLeNet neural network to detect whether there is a human face in this area,if so,records the position information of the human face,and finally passes the human face The movement distance in the vertical direction and the horizontal direction is used to judge whether it is a standing student face,so as to realize the detection and positioning of the standing face in the intelligent recording and broadcasting system.The experimental results show that the method in this paper can basically realize the function of detecting and positioning students’ standing faces.The main work of this article is as follows:(1)Train a face detection network.First,the GoogLeNet network is used to perform transfer learning on the face image to train a face detection network.(2)Identify the active area.The frame difference method and skin color detection are used to determine the active area of skin color in the classroom video,and the presence of human faces is detected in this active area,which reduces the amount of calculation and improves the efficiency of the algorithm.(3)Standing face detection and positioning.Use the trained face detection network in the active area to perform face detection and mark locations;then partition the classroom,set different thresholds for different partitions,and determine the standing face by analyzing the relationship between facial motion data and the threshold and record position.The method in this paper only performs face detection on the active area with skin color,which reduces the amount of calculation and improves the operation efficiency of the algorithm.Multiple experiments have shown that the trained GoogLeNet network can distinguish between face images and non-face images,and the detection and localization of standing students in the classroom can be basically realized.