Hand Gesture Recognition of an Individual using Model based Clustering Technique to obtain Better Accuracy Compared with Support Vector Machine

The purpose of this study is to demonstrate the accuracy of a new hand gesture recognition system based on Model Based Clustering and to compare it with a Support Vector Machine (SVM). Materials and Methods:A total of forty unique pictures of hand gestures were extracted from a dataset that was hosted on Github and made available to the public. Two groups are included in this research work. (N = 20) Model-based clustering method and (N = 20) SVM. The sample size is calculated with G power of 80 % and significance p < 0.05. Results: The innovative Model Based Clustering technique has achieved 96.66 % of accuracy and 0.46066 standard deviation while SVM has achieved 90.79 % of accuracy and 0.01861 standard deviation. The Model Based Clustering algorithm has significant accuracy of 0.000 (p < 0.05). Conclusion: The performance of the hand gesture recognition using clustering technique has provided better accuracy and reduced time complexity than Support Vector Machine (SVM).