Keyframe Extraction from Motion Capture Data by Optimal Reconstruction Error

In order to compress motion capture data in real-time,a method is proposed to extract the keyframes with an optimal reconstruction error.By the method,firstly the reconstruction error is defined to evaluate the availability of keyframe.Then a frame decimation algorithm is used to compute the reconstruction error for each decimation frame and sorted into a queue according to the error value.A reconstruction error curve is figured out through the queue and an optimal compression rate is derived from the curve.Finally,a set of keyframes are extracted from the original motion data by the compression rate.Experimental results show that this approach can be applied to effectively extract keyframes with optimal reconstruction error and meet the requirement of real-time data compression.