Face Recognition Technology And Software System Design For The Greeting Service Robot

With the development of science and technology and the aggravating trend of aging social population,the market demand of service-oriented robots is becoming more and more exuberant.So the greeting service robots,as a kind of service-oriented robots,have received extensive attention.Considering the openness and diversity of the greeting service robots' application environments,automatic face recognition system for greeting service robots is required to be intelligent,to have good robustness with regard to environmental changes and to be quick and convenient.So the face recognition algorithm and software system for the greeting service robots have become the hotspots of the current research.This paper first analyzes and summarizes the current research status of the face detection technology and the face recognition technology at home and abroad.Then the corresponding face recognition algorithm model is implemented with C++ language.On this basis,the software development of face recognition system is completed with the MFC class library and database software.The system has a reference role to the application of face recognition technology to the greeting service robots,and can meet certain practical needs.The main work of this paper can be summarized as follows:(1)In consideration of the higher requirement of speed and accuracy of face detection algorithm for the greeting service robots,this paper adopts face detection method based on the Adaboost algorithm.In view of its disadvantage of high false detection rate,this paper eliminates some of the non-face images that had been mischecked to reduce the false detection rate by detecting eyes as a two-step confirmation,and then does some experiments with it.The test results show that the face detection method can effectively reduce the false detection rate.(2)Because the application environments of the greeting service robots have many characteristics,such as complicated background and changeable lighting conditions.The quality of the collected face images is often uneven.In order to solve this problem,this paper introduces the strategy of comprehensive evaluation of face images' quality.It eliminates the low quality face images by scoring the six objective indexes of the face image to get the total score.Furthermore,for the sake of coping with the problem of the significant impact on the recognition effect due to the illumination fluctuations,the paper adopts LTP operator to extract face features from the face images.The experimental results verify that the face recognition method can effectively improve the performance of face features extraction and recognition.(3)On the basis of previous research,this paper designs and implements an automatic face recognition system software.It completes the whole process from the registration of the face information to the final automatic face recognition.It verifies the feasibility of the method using in the paper,and has laid a great groundwork for the face recognition research work afterwards.