Research And Implementation Of Sensitive Face Recognition Based On Optimized PCA Algorithm

With the development of science and academic progress,artificial intelligence these days will still be in an irreversible tendency and in an unprecedented period of prosperity that is obvious to all.Most software based on artificial intelligence in all walks of life is in the trend of times,among which face recognition is a typical application.Because of the transformation of traditional media industry to innovative integrated media and the rise trend of access to huge amounts of information,the use of the original outdated manual-content-audit mode is outdated,which requires the technical system to match the corresponding changes.In the integrated production platform of Shenzhen TV station,this project innovates the review link of the broadcast of the master control system by introducing the automatic face recognition technology,which has accumulated rich theoretical research basis in recent decades,and a large number of recognition algorithms have emerged.However,although the face recognition technology can obtain superior performance,face recognition is easy be affected by face angle,shade from other objects,and other factors.At the same time,face recognition application in the media industry is still a complex problem because of the high speed of face recognition required by live broadcast.Based on the careful analysis of the classical algorithms of face recognition,combined with the feature extraction technology of face recognition technology domestic and overseas,this paper in order to improve the efficiency of face detection and the accuracy of face recognition,PCA dimension reduction algorithm is optimized,mainly through the extraction of facial contour feature points and the update of facial feature model generation.At the same time,according to the actual application scenario of Shenzhen TV station,we use the optimized PCA detection algorithm to design and develop our own sensitive face feature library,At the same time,the recognition of feature base is tested from two aspects of functionality and system performance,and finally meets the technical requirements of the review link of the general control system.Based on the development status and trends,adding face recognition technology in the live media link not only solves the labor cost of manual review,but also improves the accuracy and efficiency of content review,which is of great value for the subsequent application of the technology to the content production link and even the development of the whole radio and television industry.