A user-specific Machine Learning approach for improving touch accuracy on mobile devices

Weir, D., Rogers, S. , Murray-Smith, R. and Löchtefeld, M. (2012) A user-specific Machine Learning approach for improving touch accuracy on mobile devices. In: 25th ACM Symposium on User Interface Software and Technology, Cambridge, MA, 7-10 Oct 2012, (doi: 10.1145/2380116.2380175)

Full text not currently available from Enlighten.

Publisher's URL: http://dl.acm.org/citation.cfm?id=2380175&CFID=303140318&CFTOKEN=48499611


We present a flexible Machine Learning approach for learn- ing user-specific touch input models to increase touch ac- curacy on mobile devices. The model is based on flexible, non-parametric Gaussian Process regression and is learned using recorded touch inputs. We demonstrate that signifi- cant touch accuracy improvements can be obtained when ei- ther raw sensor data is used as an input or when the device’s reported touch location is used as an input, with the latter marginally outperforming the former. We show that learned offset functions are highly nonlinear and user-specific and that user-specific models outperform models trained on data pooled from several users. Crucially, significant performance improvements can be obtained with a small (≈ 200) num- ber of training examples, easily obtained for a particular user through a calibration game or from keyboard entry data.

Item Type:Conference Proceedings
Additional Information:ISBN: 9781450315807
Glasgow Author(s) Enlighten ID:Murray-Smith, Professor Roderick and Rogers, Dr Simon and Weir, Mr Daryl
Authors: Weir, D., Rogers, S., Murray-Smith, R., and Löchtefeld, M.
College/School:College of Science and Engineering > School of Computing Science

University Staff: Request a correction | Enlighten Editors: Update this record