Abstract

Human shadow is one of the most important aspects which can be used as a new biometric information for gait recognition. However, human moving cast shadow extraction is still a hard task due to the effects of umbra-penumbra phenomenon, backlighting, light changing and etc. In this paper, based on the Gragh Cuts theory, a new entropy-expressed texture and colour fused extraction method for human moving cast shadow is proposed. This method built data items by the two fusing attributes mentioned above, used the gradient to get smooth items, and extracted the whole-optimized shadow region via employing the Graph Cuts theory. Some experimental results show that our method is stable and efficient under the outdoor imaging circumstances.

Paper Details