FingerInput

Project Details
Project Name :

FingerInput

Project Category :

02Body Interaction

Skill :

Best Paper Award

Client Review :

(ISS 2018)

Description :

FingerInput

Capturing Expressive Single-Hand Thumb-to-Finger Microgestures for On-Skin Input


Single-hand thumb-to-finger microgestures have shown great promise for expressive, fast and direct interactions. However, pioneering gesture recognition systems each focused on a particular subset of gestures. We are still in lack of systems that can detect the set of possible gestures to a fuller extent. In this paper, we present a consolidated design space for thumb-to-finger microgestures. Based on this design space, we present a thumb-to-finger gesture recognition system using depth sensing and convolutional neural networks. It is the first system that accurately detects the touch points between fingers as well as the finger flexion. As a result, it can detect a broader set of gestures than the existing alternatives, while also providing high-resolution information about the contact points. The system shows an average accuracy of 91% for the real-time detection of 8 demanding thumb-to-finger gesture classes. We demonstrate the potential of this technology via a set of example applications.


The publication received the Best Academic Paper Award at ACM ISS’18.


Publication

Mohamed Soliman, Franziska Mueller, Lena Hegemann, Joan Sol Roo, Christian Theobalt, Jürgen Steimle
Proceedings of the 2018 ACM International Conference on Interactive Surfaces and Spaces (ISS'18)
[Pdf] [Publisher]