Repository logo
  • English
  • Čeština
  • Deutsch
  • Español
  • Français
  • Gàidhlig
  • Italiano
  • Latviešu
  • Magyar
  • Nederlands
  • Português
  • Português do Brasil
  • Suomi
  • Svenska
  • Türkçe
  • Қазақ
  • বাংলা
  • हिंदी
  • Ελληνικά
  • Log In
    or
    New user? Click here to register.Have you forgotten your password?
Repository logo
  • Communities & Collections
  • Research Outputs
  • Projects
  • People
  • Statistics
  • English
  • Čeština
  • Deutsch
  • Español
  • Français
  • Gàidhlig
  • Italiano
  • Latviešu
  • Magyar
  • Nederlands
  • Português
  • Português do Brasil
  • Suomi
  • Svenska
  • Türkçe
  • Қазақ
  • বাংলা
  • हिंदी
  • Ελληνικά
  • Log In
    or
    New user? Click here to register.Have you forgotten your password?
  1. Home
  2. Scholalry Output
  3. Publications
  4. Lightweight Deep Multiscale Feature Learning for Improved Human Activity Recognition
 
  • Details
Options

Lightweight Deep Multiscale Feature Learning for Improved Human Activity Recognition

Journal
2024 IEEE Region 10 Symposium (TENSYMP)
Date Issued
2024
Author(s)
Baruri Sai Avinash
Abhishek Yadav
Kumar, Binod 
Department of Electrical Engineering 
DOI
10.1109/TENSYMP61132.2024.10752181
Abstract
Human Activity Recognition (HAR) is a significant field within human-computer interaction, often utilizing deep learning (DL) techniques for improved performance. However, DL's resource-intensive nature poses challenges, especially on memory-limited devices like smartphones and other edge devices. To address this issue, we propose a novel approach combining Wavelet Transform (WT) and DL. By decomposing smartphone signals using WT and feeding them into a Multiscale Con-volutional Neural Network (MSCNN), our method effectively extracts features at different scales for adaptive HAR without the need for manual feature extraction. This WT-MSCNN model offers advantages in processing non-stationary data, extracting HAR information effectively, and maintaining robustness to scale uncertainties in signals. Our experimental results indicate that the proposed methodology attains an accuracy of 98 %, a sensitivity of 99%, a specificity of 97%, and an Fl-score of 98.10% on NVIDIA Tesla T4 GPU (Graphics Processing Unit). Additionally, our model has demonstrated enhanced throughput, outperforming fine-tuned MobileNet-V3 and SqueezeNet Models by 83.59% and 25.72% respectively on the Raspberry Pi 4 hardware platform.
Subjects
  • Deep Learning

  • Graphics Processing U...

  • Human Activity Recogn...

  • Multiscale Convolutio...

  • Raspberry Pi 4

  • Wavelet Transform

Copyright © 2016-2025  Indian Institute of Technology Jodhpur

Developed and maintained by Dr. Kamlesh Patel and Mr. C. Chhatwani, S. R. Ranganathan Learning Hub, IIT Jodhpur.

Built with DSpace-CRIS software - Extension maintained and optimized by 4Science

  • Cookie settings
  • Privacy policy
  • End User Agreement
  • Send Feedback