ConvNet on Doppler Spectrogram: A Novel Enactment of Hand Gesture Recognition

5 Pages Posted: 29 Aug 2019

See all articles by Robin

Robin

College of Engineering, Trivandrum

Jeena

College of Engineering, Trivandrum

Date Written: August 28, 2019

Abstract

Consumer electronics industry is investing much of its R&D backbone in realising the best human-machine interface at any point of time. Out of the different technologies conceived to meet this demand, hand gesture recognition is becoming a widely accepted one. Most of the existing systems are build by adding an additional sensing element to the device which is to be controlled. The present work introduces an external device free hand gesture recognition system using convolutional neural networks. The concept is proved in MacBook Air 13 inch laptop by converting it into an active sonar system and it can be extended to any of the similar mobile device with a pair of microphone and speaker. The system is able to decode and classify four types of hand gestures with a classification accuracy of 90%.

Keywords: Spectrogram, Convolution Neural Network, Doppler, Hand Gesture

Suggested Citation

V T, Robin and R S, Jeena, ConvNet on Doppler Spectrogram: A Novel Enactment of Hand Gesture Recognition (August 28, 2019). In proceedings of the International Conference on Systems, Energy & Environment (ICSEE) 2019, GCE Kannur, Kerala, July 2019. Available at SSRN: https://ssrn.com/abstract=3444033 or http://dx.doi.org/10.2139/ssrn.3444033

Robin V T (Contact Author)

College of Engineering, Trivandrum ( email )

Thiruvananthapuram, Kerala 695016
India

Jeena R S

College of Engineering, Trivandrum ( email )

Thiruvananthapuram, Kerala 695016
India

Register to save articles to
your library

Register

Paper statistics

Downloads
9
Abstract Views
87
PlumX Metrics