Recent methods and databases in vision-based hand gesture recognition

摘要: The paper surveys RGB and RGB-D sensors based hand gesture recognition methods.Dynamic as well static (posture/pose) methods are reviewed.Qualitative quantitative comparison of algorithms is provided.Twenty-six publicly available gesture/posture databases also reviewed.Discussion on unresolved issues future research directions provided. Successful efforts in within the last two decades paved path for natural human-computer interaction systems. Unresolved challenges such reliable identification gesturing phase, sensitivity to size, shape, speed variations, due occlusion keep still very active. We provide a review vision-based reported 16 years. using cameras reviewed with qualitative comparisons algorithms. Quantitative done set 13 measures chosen from different attributes algorithm experimental methodology adopted evaluation. point out need considering these together accuracy predict its success real-world applications. reviews 26 provides web-links their download.