A 3D-touch interface by using EMG

Abstract

The 2D-Touch interaction for users to provide on-screen position values as user input provides various input methods such as Touch, Long-Touch, Drag and so on. However, it does not provide a way for the user to provide input through control of the force. To achieve this, 3D-Touch interface came out. This is a new way of interacting that can be used by adding force values as input in 2D-Touch. 3DTouch adds depth of force to add the strength z of force to the coordinates of the existing position input x, y. Therefore, it is possible to diversify the input by force input in an environment where 2D-Touch is not possible, such as constrained space. However, 3D-Touch has the disadvantage that it can be measured only by a screen device capable of measuring force. For this reason, the 3D-Touch interface is not popularly used and is used only in limited products. Another way to measure force is through electromyogram, EMG. EMG (surface electromyography signal) is a biological signal used to sense the degree of activation and mobilization patterns of the nerve roots that are regulated by the nervous system during muscle. The EMG signal can cause a change in the signal to distinguish the 3D-Touch from the 2D-Touch by the applied force. Therefore, if users use a device that can measure EMG signals, they can use 3D-Touch interaction on devices that do not provide a 3D-Touch screen. In this paper, we test whether the three input methods of Touch, Peek, and Pop can be classified into EMG signals. CNN (Convolutional Neural Networks) is used to distinguish the EMG signals of each input. © 2019 Association for Computing Machinery.