Motion Behavior Feature Segmentation Based on Intelligent Vision
Zhao Meihuan · 2020 IEEE International Conference on Power, Intelligent Computing and Systems (ICPICS) · 2020
In order to improve the ability of 3D virtual reconstruction and recognition of motion behavior features, human motion behavior features are extracted and segmented. A motion behavior feature segmentation algorithm based on intelligent vision analysis is proposed. The motion behavior feature image is collected and analyzed under multimedia vision, and multi-scale Retinex corner selection is carried out for the acquired motion behavior feature image. The method of 3D dynamic tracking recognition is used to analyze the dynamic image of motion behavior feature, and the histogram distribution feature fuzzy learning method is used to realize the 3D dynamic tracking recognition of motion behavior feature. The motion behavior feature segmentation is realized by analyzing the dynamic characteristics of the motion behavior feature combined with the three-dimensional motion manifold feature. The simulation results show that, this method can improve the intelligent judgment and detection ability of the moving image, and the feature expression ability of image segmentation is strong, and the performance of template matching is better.