Adaptive robust moving hands recognition under complex lighting condition

This paper describes an adaptive robust moving hands recognition algorithm using Kinect V2, which can detect the bare hands or hands with gloves in the real-time image stream under complex lighting condition. Firstly, according to the Bayes criterion, a novel skin color classification is built on the best separation plane in color space, which is found through linear discriminant analysis (LDA). Secondly, an adaptive learning rate and connected component theory are added to the traditional background subtraction. Finally, this new background subtraction and LDA skin color classification are combined together with an adaptive updated skin color look-up-table. In experiment results, this algorithm presents a satisfactory performance in hand detection under complex lighting condition.