Extracting features of fingertips bending by using self-organizing map

Abstract

In this paper the method of Self-Organizing Maps (SOM) is introduced to analyze the human grasping activities of human fingertips bending using the low cost DataGlove called as GloveMAP. The research shows that the proposed approaches capable to utilize the effectiveness of the SOM for creating the grasping features of the bottle object. After the iterative learning of net-trained, all data of the trained network will be simulated and finally self-organized. The final result of the research study shows the fingertips features extraction were generated from the several grasping activities and verify the validity of the analysis through simulation with human grasp data captured by a GloveMAP.