In order to enhance the social integration ability of blind individuals and enhance their connection with society,the study utilizes an improved YOLOv3 scene understanding algorithm and Kinect sensor to design a music interactive landscape experi-ence system.Firstly,the YOLOv3 algorithm is improved using clustering methods to enhance its recognition and classification capa-bilities;Then,the YOLOv3 algorithm is fused with Kinect sensors to construct music interaction designs;Finally,use comparative methods and practical applications to verify the performance of the music interaction design system.The results indicate that YOLOv3+Kinect outperforms the comparison method in terms of system parameter structure.Its loss value,data recall rate,mAP％,and aver-age data testing time are 21.58％,15.13％,32.7％,4.37％,43.86％,34.21％,0.98 seconds,and 0.15 seconds higher than the comparison method,respectively.This indicates that the system constructed in the study has high feasibility and practicality in the de-sign of music interaction landscapes,providing a brand new music interaction experience for blind groups.
维普
万方数据