Skip to Main Content
In this paper, we solve the searching problem by high level features used by sign language recognition. Firstly, we find the face in video frames that has complex background, and then we find the left sign and right sign in specific areas. By computing the signs' length, position, velocity, acceleration, Fourier figure descriptor and etc, we generate the signs' dynamic features. Consequently, we segment the video frames by motion features. As for each segment, we generate a HMM. When a clip of sign language inputs, we also get the feature serials, and then we compare the possibility of the input serials in each HMM. Experiment results on a large of sign language videos show that our searching system performs much better than existing methods on sign language video searching systems. Compared with the traditional methods, our system reduces the average searching time by half and the retrieval precision has doubled.