Objects Classification based on Hand Grasping in Virtual Reality Environment

Recent advancements in Artificial Intelligence and machine learning methods have been the focus of much research in different application domains due to their possibility to enable intelligent tasks. Integrating intelligence for human-machine interaction is an interesting topic that may improve the quality of life when used for robotics, prosthetics, and rehabilitation domains, to name a few. This work presents the analysis of the touch interaction between hands and objects at the moment of grasp. A virtual reality environment has been employed to collect the dataset in order to apply machine learning methods. Three different algorithms have been adopted to recognize the touched object achieving a classification accuracy of 94.7% for the KNN and SVM, and 98.2% for the LSTM network.