A novel key frames matching approach for human locomotion interpolation

A novel key frames matching approach for human locomotion interpolation is proposed in this paper. Firstly, locomotion data is transformed from body coordinate system to world coordinate system based on recursive algorithm; secondly, features of two feet and step distance are extracted; thirdly, key frames are extracted based on feature curve and the locomotion sequence is segmented; fourthly, a new approach for key frames matching is put forward; finally, transition frames are interpolated between the matched key frames using quaternion Slerp (Spherical linear interpolation) algorithm and linear interpolation algorithm. The main contributions of this paper are: i) different locomotion sequences can be synthesized in a controllable manner with good effect; ii) to extract locomotion features conveniently, locomotion data in body coordinate system is transformed to world coordinate system; iii) to analyze locomotion sequences intuitively and accurately, key frames are extracted and locomotion sequences are segmented based on feature curve analysis; iv) to avoid illogical jump of traditional interpolation between key frames, a new key frames matching strategy is put forward and only the key frames between two continuous segments can be chosen as matching frames. Experimental results based on the CMU MoCap database show that the proposed method can confirm the logical correctness and naturalness in the junction of two locomotion sequences.