Labanotation is a widely used dance recording system, which plays an important role in inheriting and protecting folk dances. However, manual drawing of dance notation is time-consuming and labor-intensive. Therefore, research on the automatic generation of Labanotation has received extensive attention. In this article, a method to automatically generate Labanotation based on motion capture data is proposed. Firstly, in order to facilitate the calculation of the bending angle and motion feature sequences, the BVH-formatted Euler angle data is converted into a sequence of the 3D Cartesian world coordinates according to the analysis of human bones. Secondly, a motion segmentation method based on kinematics features and rhythms is proposed, which can segment different styles of motion into fragments for further identification. Finally, we generate the Labanotation with the platform we developed. Experiments show that our segmentation method has good accuracy in segmenting action sequences, which makes the generated Labanotation more reliable.

中文翻译：

---

基于节奏的舞蹈动作分割与标注方法

Labanotation是一种广泛使用的舞蹈记录系统，在继承和保护民间舞蹈方面起着重要作用。然而，手工绘制舞蹈符号是费时且费力的。因此，关于自动标注的研究受到了广泛的关注。本文提出了一种基于运动捕捉数据自动生成Labanotation的方法。首先，为了便于计算弯曲角度和运动特征序列，根据人体骨骼的分析，将BVH格式的欧拉角数据转换为3D笛卡尔世界坐标的序列。其次，提出了一种基于运动学特征和节奏的运动分割方法，该方法可以将不同风格的运动分割为多个片段，以进行进一步的识别。最后，我们使用我们开发的平台生成Labanotation。实验表明，我们的分割方法对动作序列进行分割具有很好的准确性，使得生成的Labanotation更加可靠。