In order to improve the recognition and optimization effect of dance movements, this paper combines multi-threshold image segmentation technology to perform intelligent recognition of dance movements. Moreover, this paper analyzes the window focus calibration process in the cascade multi-threshold expansion method and focuses the window on the object body to expand from small to large. Then, due to the local optimum problem of single-layer expansion, this paper proposes a cascade expansion method, and comparative experiments analyze the effect of this method on window optimization. In addition, this paper combines the actual situation of dance movement training to construct an intelligent action recognition and dance movement optimization system based on multi-threshold image segmentation. The research shows that the intelligent action recognition and dance motion optimization method based on multi-threshold image segmentation proposed in this paper has a good effect.

中文翻译：

---

基于多阈值图像分割的智能动作识别与舞蹈动作优化

为了提高舞蹈动作的识别和优化效果，本文结合多阈值图像分割技术对舞蹈动作进行智能识别。此外，本文分析了级联多阈值扩展方法中的窗口焦点校准过程，将窗口聚焦在对象主体上，由小到大扩展。然后，针对单层扩展的局部最优问题，提出了一种级联扩展方法，并通过对比实验分析了该方法对窗口优化的效果。此外，本文结合舞蹈动作训练的实际情况，构建了基于多阈值图像分割的智能动作识别和舞蹈动作优化系统。