Hand Structure Analysis for Finger Identification and Joints Localization

Abstract

The development of kinematic model of hand can play
a vital role in hand gesture recognition and Human
Computer Interaction (HCI) applications. This paper
proposes an algorithm for finger identification and joints
localization, thus generating the kinematic model of human
hand by means of image processing techniques. Skin tone
analysis and background subtraction is carried out for hand
detection in the workspace. Geometric features of hand are
used for hand identification (left or right), finger
identification and joints localization. Proposed algorithm is
tested for diverse hand poses and remarkable results are
produced. Algorithm not only generates the kinematic model
for the different orientations of the hand but also have very
low computational cost.