Research On Adaptive Methods And Techniques For Athletic Gaze Analysis Model

"The eye is the window to the psyche." Recent psychological and visual research has shown that human mental activity can be expressed to some extent by gaze.However,the analysis of mental activity by gaze also needs to be integrated with serve mechanics and tactics,and also needs to distinguish scenarios in which the character’s gaze features are found.In this paper,we investigate the adaptive methods and techniques of sports gaze analysis models for the adaptive problems of image extraction in gaze acquisition,the adaptive problems of key figure determination,and the correlation between the gaze features of sports figures and the gaze target.In order to study the gaze of moving characters in a targeted way,the paper divides the gaze into "motion gaze" and "non-motion gaze",and gives the definition of motion gaze.In order to analyze the motion gaze,we focus on the methods and techniques of longitudinal and lateral motion gaze acquisition,propose the relevant algorithm for motion gaze acquisition,and give the correlation analysis method of motion gaze with the design of table tennis player’s serve technique and tactics.The adaptive problem of image extraction is studied,and the adaptive method of motion video image extraction is given.The adaptive algorithm of image extraction based on artificial neuron network is proposed,which is realized by the mutual cooperation of decision module,selection module and training module.The training module uses the best interval frames provided by the selection module and the video attribute data provided by the decision module to train the network model of the best interval frames;the adaptive problem of determining key people in motion video is studied,and the adaptive method of key face selection in motion video is given,and the adaptive algorithm of key face recognition based on face features and visual distance is proposed.adaptive recognition algorithm,which uses the triangle similarity principle to screen out the face features with the closest visual distance in each picture,then selects the key face features in a set of pictures by face feature clustering method,and finally calculates the similarity between the face features and the key face features in each picture in turn to screen out the key faces in each picture;the experiments verify the effectiveness of the above method.The legacy software for gaze acquisition and analysis was reconstructed to add the above research results,and the software has been submitted to the national table tennis team for trial.