Skip to Main Content
A novel method to recognize human activities based on videos is proposed in this paper. These videos are captured by a camera mounted to a human body. We can estimate the activity from the changes of scenes in videos. In this paper, we use the improved diamond search block-matching method to calculate the motion vector. Then we extract key information from the motion vector filed, and design a feature descriptor to describe the motion in frames in a video which can distinguish different motions. After getting feature descriptors, we use SVM classifier to classify different motions with a machine learning method. Experimental results show that our method successfully identifies simple motion such as walking, running, going upstairs and going downstairs. And the block size and the frequency of videos have impacts on classification precision.