Robust Human Motion Recognition Based on Multi-directional Motion Representation

A robust human motion recognition method is proposed in this paper. The basic motion history image (MHI) has been updated considering multidimensional history and energy images to calculate the feature vectors for recognizing various complex human motions. The ´rajio-taiso´ (Japanese Radio Exercise) has been recognized by employing the method. This exercise was conducted by some non-professionals and hence the input video data has lots of disarray and variations along with the complex movements of the exercise. The recognition rate has been satisfactory in spite of the complex nature of the input data, whereas the basic motion history image was performed with the video data of aerobics exercise, performed by professionals and formal environment.