Unsupervised feature learning for human activity recognition

To solve the problems that human limitations may cause the loss of important information, thus affecting the classification results, a feature extraction method based on unsupervised feature learning techniques was proposed. Unsupervised feature learning method to learn multiple feature maps was used and concatenated together. This method can avoid the loss of important information, and also can significantly reduce the scale of unsupervised feature learning model used. To evaluate the proposed method, experiments on a public human activity recognition dataset were performed, using three commonly used unsupervised feature learning models, and finally using support vector machines to classify activities. The results show that the proposed feature extraction method achieves good results, and has certain advantages compared with other methods. ©, 2015, National University of Defense Technology. All right reserved.