Hand Posture Recognition in 3D space using Ensemble Voting Classifier

Among different hot topic research areas, hand gesture recognition is considerably important and applicable one. This isfrequently usable because of the application of this in human computer interaction, like as wearable gadgets, driving cars,automated robots, biometric authentication, and health area. Furthermore, this task is very applicable in vision systems, signlanguage study etc. Hand posture recognition could be addressed in real-time and offline modes. As in most of the cases, thealgorithms for offline causes of the problem could also be applied to online mode. In this paper, an ensemble-voting classifier isutilized for the offline recognition of hand posture. Ensemble models that reflect the ideas of base classifiers and aggregate theirresults properly, achieve significantly better results than single classifiers. The experimental comparison of the proposed methodwith previous methods demonstrates that the proposed method performs better. Our method achieves the accuracy and BalancedError Rate (BER) of 98.39% and 0.016 respectively.

کلیدواژه ها:

Human computer interaction ، 3D hand posture recognition ، Posture recognition ، Ensemble classifier