Human hand image analysis extracting finger coordinates using circular scanning

This paper presents a Circular scan fingertip detection algorithm for accurate calculation of fingertip coordinates based on a hand image captured from a depth sensor. It properly analyzes both straight and bent fingers and is rotation invariant. The algorithm can be used for development of user interfaces based on human gesture analysis, such as a multi-modal user interface TouchTable. Advantages of the proposed algorithm over other methods of fingertip detection are described. The paper also introduces DepthTip, a public dataset that can be used to evaluate accuracy of fingertip detection algorithms.