RadarNet: Efficient Gesture Recognition Technique Utilizing a Miniature Radar Sensor
Authors:
Open Access
- 6 May 2021
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-14
Abstract
Gestures are a promising candidate as an input modality for ambient computing where conventional input modalities such as touchscreens are not available. Existing works have focused on gesture recognition using image sensors. However, their cost, high battery consumption, and privacy concerns made cameras challenging as an always-on solution. This paper introduces an efficient gesture recognition technique using a miniaturized 60 GHz radar sensor. The technique recognizes four directional swipes and an omni-swipe using a radar chip (6.5 × 5.0 mm) integrated into a mobile phone. We developed a convolutional neural network model efficient enough for battery powered and computationally constrained processors. Its model size and inference time is less than 1/5000 compared to an existing gesture recognition technique using radar. Our evaluations with large scale datasets consisting of 558,000 gesture samples and 3,920,000 negative samples demonstrated our algorithm’s efficiency, robustness, and readiness to be deployed outside of research laboratories.Keywords
This publication has 14 references indexed in Scilit:
- Interacting with Soli2016
- A Highly Integrated 60 GHz 6-Channel Transceiver With Antenna in Package for Smart Sensing and Short-Range CommunicationsIEEE Journal of Solid-State Circuits, 2016
- SoliACM Transactions on Graphics, 2016
- Deep Residual Learning for Image Recognition2016
- Accurate, Robust, and Flexible Real-time Hand Tracking2015
- SoundWave2012
- Humantenna2012
- SideSight2008
- Gestural and audio metaphors as a means of control for mobile devices2002
- “Put-that-there”1980