A motion analysis and classification based on PCA and ELM classifier

In this paper, we propose a method to classify K-pop dance based on motion data obtained from Kinect V2 for research of motion classification and development of anti-plagiarism system. To do this, 200-point dances of K-pop are acquired. Dance motions from 40 amateur dancers are acquired to construct a total of 400 data. The proposed classification method consists of three steps. First, we obtain 13 significant angles that represent dance motion well in each frame, perform preprocessing process, obtain four statistical angles every 30 frame of each angle, and connect them together to generate feature vector. Second, PCA (Principal Component Analysis) is used to reduce the dimension of the feature vector. Finally, we design an ELMC (Extreme Learning Machine Classifier) that uses dimensionally reduced feature vectors of dance motion as input data. ELM (Extreme Learning Machine) can be learned fast because it calculates final weights only with one forward pass. As a result of comparing the performance of the previous method without statistical angles and the method using KNN and SVM classifier, classification rate of the proposed method is better than the previous method.