基于骨骼信息的原生态舞蹈数字化保护研究  

Research on digital protection of primitive dance based on skeletal information

作　　者：谢静文 王卫星[1,2] 马连丹[2] 胡宁峰 Xie Jingwen;Wang Weixing;Ma Liandan;Hu Ningfeng(Key Laboratory of Modern Manufacturing Technology of Ministry of Education,Guizhou University,Guiyang,Guizhou 550025,China;School of Mechanical Engineering,Guizhou University)

机构地区：[1]贵州大学现代制造技术教育部重点实验室,贵州贵阳550025 [2]贵州大学机械工程学院

出　　处：《计算机时代》2023年第9期136-141,共6页Computer Era

基　　金：贵州省科学技术基金(黔科合基础[2020]1Y262);贵州省教育厅青年科技人才成长项目(黔教合KY字[2018]112)。

摘　　要：少数民族原生态舞蹈是国家非物质文化遗产的重要组成部分,但由于文化的多元化使其发展和传承受到冲击。为了更好地保护与传承原生态民族舞蹈,本研究结合Kinect深度相机设计并开发了一套原生态舞蹈动作的学习框架。该框架通过Kinect采集并存储专业原生态舞者动作的骨骼信息,并使用改进的3D CNNS模型对原生态舞蹈进行识别,根据识别结果,采用Cosine-DTW算法对训练者的舞蹈动作进行评价,从而实现原生态舞蹈动作的数字化保护与传承。The primitive dance of ethnic minorities is an important part of the national intangible cultural heritage,but its development and inheritance are affected by the diversity of culture.In order to better protect and inherit the primitive dance,a learning framework of primitive dance movements is designed and developed combined with the Kinect depth camera.The skeletal information of professional primitive dancers'movements is collected and stored by Kinect,and the improved 3D CNNS model is used to recognize the primitive dance.Based on the recognition results,the Cosine-DTW algorithm is used to evaluate the dance movements of the trainers,so as to realize the digital protection and heritage of primitive dance movements.

关 键 词：KINECT 骨骼信息 3D CNNS Cosine-DTW 原生态舞蹈 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]