A Study on Grasp Recognition Independent of Users' Situations Using Built-in Sensors of Smartphones
Authors:
Open Access
- 6 November 2015
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 69-70
Abstract
There are many hand postures of smartphone according to the users? situations. In order to support appropriate inter-face, it is important to know user?s hand posture. To recognize grasp postures, which is not depend on users? situations, we consider using smartphone?s touchscreen and their built-in gyroscope and accelerometer and use support vector machine (SVM). In order to evaluate our system, we described the result of the experiments when users are using the devices in the room and on the train. We knew that our system could be feasible for personal use only system by improving the information from the accelerometer. We also collected users? data when users are sitting in the room. Results showed that grasp recognition accuracy under 5 and 4 hand postures were 87.7%, 92.4% respectively when training and testing on 6 users.Keywords
Funding Information
- The Japan Society for the Promotion of Science (25330227)
This publication has 5 references indexed in Scilit:
- iRotateGrasp2013
- IrotateGrasp2013
- GripSense2012
- LIBSVMACM Transactions on Intelligent Systems and Technology, 2011
- What shall we teach our pants?2000