Dates
Thursday, January 20, 2022 - 02:30pm to Thursday, January 20, 2022 - 04:00pm
Location
Zoom (contact events@cs.stonybrook.edu for more details)
Event Description

Abstract: Over the past decades, mobile computing technologies developed rapidly with the wide adoption of mobile devices. As the dominant input modality in mobile computing, touch input has been extensively investigated by a considerable amount of research and has various real-world applications. Modeling touch pointing is essential to developing touchscreen interfaces and their applications. In this report, we first explore existing models and theories of touch pointing. Then, we propose the Rotational Dual Gaussian Model (RDGM), a refinement and generalization of the Dual Gaussian model that accounts for the finger movement direction, and present its applications on touchscreen keyboard decoding. We show that the RDGM reduces target selection error rate, more accurately predicts touch point distributions, and improves soft keyboard decoding accuracy. Next, we design a deep learning model trained on touch screen typing input to detect motor impairments in early Parkinson's Disease (PD) of older adults. Using this deep learning method shows promising results and outperforms the previously used SVM-based method.

Event Title
Ph.D Research Proficiency Presentation: Yan Ma, 'Touch Pointing Models and Their Applications'