Design And Development Of DSP-based Face Recognition System For Rehabilitation Nursing Robot Bed

At present,patients still need to operate the nursing beds on sales by themselves or under the assistance of families and medical staffs,however,such nursing beds can' t meet the demand of future development of intelligent rehabilitation nursing industry.With the help of expert system reaso ning machine working on patient's expression states and vital signs parameters,the system can automatically analyze the patients‘ physical needs,send commands through the serial port to drive the bed to work while no one is looking after the patients,if the face recognition system based on DSP is successfully added to the Rehabilitation Nursing Robot bed.This paper discusses and studies the realization of the DSP-based face recognition system,which includes face detection,face recognition,expression recognition and expert system,etc.The specific contents of the work are as followed:(1)According to the overall analysis of the face recognition system based on DSP to realize the function,put forward to complete the system design ideas,,and complete the overall scheme design of face recognition system based on DSP.(2)This paper studies the Adaboost algorithm and the eigenface method,realized face detection classifier training process,using the eigenface representation of face process,online face detection process and online face recognition process,and the face detection and face recognition algorithm transplanted to DSP in face recognition system.(3)In this paper,considering the actual application of the system,the face recognition algorithm is improved,a feature extraction method of face geometric distribution rules and Harris corner algorithm combined,according to the geometric distribution rules about face segmentation of eye and mouth area,and then the Harris corner algorithm to determine the specific location and shape of eyes and mouth.This method greatly reduces the computation and ensures the real-time performance of the system.(4)In the process of establishing expression recognition expert system rule base,the expression state and vital signs parameters are used as the inference front of expert system,and then establish the knowledge base according to the existing knowledge,case and expert experience,and the reasoning mechanism based on rule reasoning,case-based reasoning and fuzzy reasoning is used.(5)In this paper,a face recognition system based on DSP is built,the DSP integrated development environment CCS complete image acquisition,image processing,expert systems,image display and DSP and ARM control system between the serial communication tasks.The system uses CCD camera to capture images in real time,after image processing,can get the expression state,combined with the parameters of vital signs,as the expert system reasoning before,evaluate the real-time needs of patients,and send control instructions,rehabilitation nursing bed robot to complete the corresponding function.In the early stage of algorithm verification,the face detection,face recognition and facial expression recognition algorithm are tested by ORL face database and local face database respectively.The experimental results show that the algorithm has high accuracy and real-time performance.Finally,the system is connected with the rehabilitation nursing robot bed,and the whole system is tested.Based on the successful application of DSP face recognition system in Rehabilitation N ursing Robot bed,it realizes the intelligent and humanization of Rehabilitation N ursing Robot bed,and has high social popularization value.