Feature Extraction and Segmentation Techniques in a Static Hand Gesture Recognition System
Subhamoy Chatterjee, Piyush Bhandari, Mahesh Kumar H. Kolekar · 2017
In this chapter, a skin color region segmentation method based on K-means clustering and Mahalanobish distance is presented for static hand gesture recognition. The proposed segmentation method is robust to illumination and skin color complexion. Various shape- and contour-basedfeature extraction techniques like moment features, contour signature, and localized contour sequence features have been discussed, and the problem associated with similar-shape gesture misclassification has been pointed out. Feature enhancement methods such as F-ratio-based weighted feature extraction technique and feature fusion methods are proposed. Zoning-based shape feature extraction methods have been proposed to overcome the similar-shape gesture misclassification problem. The proposed method has shown a significant improvement in user-independent static hand gesture recognition.