Comparison of GLCM based Hand Gesture Recognition Systems using Multiple Classifiers
Maryam Naveed, Quratulain Quratulain, Arslan Shaukat · 2021
Automatic human gesture recognition system serves the purpose of human-computer interaction (HCI). In this paper, we propose a hand gesture recognition method based on multiple classifiers through feature extraction. For this, we have employed two datasets of hand gesture images. On images, different pre-processing, segmentation and morphological operations have been performed. Gray Level Co-occurrence Matrix (GLCM) based statistical features are then extracted. The hand gesture recognition is then performed by using KNN classifier, decision trees (DT), and Support Vector Machines (SVM). The proposed methodology is effective as it gives high recognition accuracies on both datasets. This study also pays heed to compare the accuracy of the three machine learning methods.