Application of AI Recognition of Human Skeletal and Joint Actions in Immersive Experience Design
Due to the diversity and complexity of human actions,existing methods can not accurately separate the human body from the background,and fail to capture key joint features,resulting in poor motion following accuracy between human actions and virtual objects.Therefore,the application research of AI recognition of human skeletal and joint actions in immersive expe-rience design is proposed.The Kinect sensor is used to obtain human motion images,threshold segmentation technology is used to segment the human body from the background environment,and extract key features of human skeletal and joints.Using features as input,BP neural network is used to achieve human skeletal and joint action recognition.Human action data are mapped to virtual objects,enabling real-time synchronization of their actions and achieving an immersive experience.The re-sults show that the Kappa coefficient of proposed method is close to 1,which has high consistency in recognition results.In the immersive interaction test,the lowest correlation coefficient of joint action angle between real human body and virtual character is only 0.96,indicating good application effectiveness.
human skeletal and joint actionAI recognitionimmersive experience
万方数据
维普
