coloriz / touch-extrapolation-mlp

Developing the neural networks that extrapolate the movement of a finger on a touchscreen.

Touch Extrapolation MLP

The dataset to train neural networks which extrapolate the movement of a finger on a touchscreen was collected by Henze et al.[1-2]. You can find more details on their papers or repository.

Model Diagram

Result

The experiments were conducted on varying window sizes. Our experiments showed that window size of 11 gives the lowest prediction error.

window size	MSE	RMSE
3	160.80	12.68
4	20.96	4.58
5	17.46	4.18
6	0.37	0.61
7	1.60	1.26
8	0.55	0.74
9	0.07	0.27
10	0.12	0.34
11	0.04	0.20
12	0.54	0.74
13	0.03	0.16

Prerequisites

• tensorflow r1.13
• matplotlib
• numpy

References

[1] Niels Henze, Markus Funk, Alireza Sahami Shirazi: Software-reduced touchscreen latency. Proceedings of the International Conference on Human-Computer Interaction with Mobile Devices and Services, 2016.

[2] Niels Henze, Huy Viet Le, Sven Mayer, Valentin Schwind: Improving Software-Reduced Touchscreen Latency. Adjunct Proceedings of the International Conference on Human-Computer Interaction with Mobile Devices and Services, 2017.