A fingertips detection method based on the combination of centroid and Harris corner algorithm

摘要

Fingertip detection is an important research topic, which contributes to silent communication. To ensure consistent performance under unconstrained environments, we propose a robust and novel method to identify fingertips of hand palm. Briefly, two-dimensional Gaussian probability distribution in YCbCr color space is introduced for segmenting hand region from background. Then feature points are detected by Harris corner detection in the binary image. With the prior information of feature points, it takes a global optimization approach of centriod and convex hull analysis to locate fingertips. Experiment results have demonstrated that the proposed system performs well because it can well detect varies of hands under different environments.