Design And Implementation Of Genealogy Platform Based On Automatic Segmentation

With the development of the economy,some traditional dance resources are faced with the situation of disappearance.How to accurately record and save them is an imp-ortant research topic.Labanotation can effectively represent all the human movements and it is easy to read and save.However,manual recording Labanotation requires a long time.Therefore,it is necessary to develop a technology for recording and saving the traditional dance movements accurately and quickly.The main work of this paper is that we propose an automatic Labanotation gener-ation algorithm converting motion capture data to Labanotation.For upper limbs,this paper presents an automatic generation algorithm based on space division.For lower limbs,this paper introduces a segmentation algorithm based on Gaussian mixture model and then convert the human motion data into Labanotation.What's more,the platform has the function of artificial Laban editing which can improve the recognition rate.The main work and innovation are as follows:(1)This paper introduces a segmentation algorithm based on Gaussian mixture model.First,We convert the human motion data into the coordinates in the world coorclinate system.Then we calculate the angular velocity of the main joints of the human body and divide the data by setting a fixed threshold.We use the Gaussian mixture model for further segmentation and find the segmentation point of the human motion data.Then we split human motion data into segments of motion.Finally,we judge the corresponding Laban symbol by judging the space position of the motion segments.(2)According to the records of Labanotation upper nodes,this paper presents an automatic generation algoritrhm based on space division.First,each frame is converted into a symbol sequence by space law and a particularly dense motion sequence is obtained.Then we use wavelet analysis to get the minimum motion rhythm and then use the classifier to decide whether every symbol is retained or not and get the final result.(3)On the basis of the frame of human motion capture data analysis,this paper adds an artificial spectrum modification function.This paper uses an optical motion capture equipment,the first step we establish the human motion model,the second step is to collect human motion data,Human motion data is stored by BVH(Biovision Hierarchical Data)file.This paper realizes the automatic data BVH data into Labanotion analysis system,in order to increase the practicability of the systcem,the design and implementation of Labanotation artificial write spectrum based on the modified spectrum platform automatic conversion system.This paper also realizes the artificial and automatic generation of Labanotation written notation for error correction function.According to our research this platform is the first Labanotion writting platform in our country.