Back to search

University of Strathclyde

Deep learning and sEMG signals based human hand gesture recognition

Abstract

dc:description.abstract

Hand prosthesis have helped amputees to partially live normal lives as healthy peoplefor ages since invented. However, due to the technical limitations such as the essentialtime required for signal sending and receiving, the state-of-the-art hand prosthesis stillcannot fully restore real hand functions. As Artificial Intelligence (AI) technologies,especially deep learning and artificial neural networks, nowadays show an impressiveperformance when building smart machines, it is possible to use them to bringimprovements to conventional hand prosthesis. This advanced AI prosthesis can learnreal hand functions through self-learning and eventually, fully achieve all real handfunctions. Moreover, with the help of current biologically inspired neural networkmodels and spiking neural networks (SNN), the power consumption and reaction delaysof a prosthesis can be further minimized.A novel approach which attempts to address the long-existing problems such asfrequent misclassification faced by current hand prosthesis is presented in this thesis.Through converting the raw surface electromyograph (sEMG) signals from amputeeswith different hand amputation levels and able-bodied people into heatmaps, with anapplied properly designed convolution neural network which extracts and learns thefeatures contained within the heatmaps, the mis-trigger disadvantage rapidly decreases.The experimental results, from 8 hand gestures so far, indicate that this novel approachis effective. Moreover, the relationship between the number of sensors, used to recordthe sEMG signals, and the recognition accuracy is also examined and presented in thisthesis. According to the experimental results, the optimal or required number of sensorswhen recording sEMG signals can be minimized for different hand gestures withoutclassification accuracy degradation.This thesis also contains another novelty: a spiking sEMG signal maps-based handgesture classification algorithm. The algorithm employs a spiking neural network as theclassifier, as well as a common heatmap technique which efficiently converts the rawsEMG signals into common heatmaps that can be used by the spiking neural network.The classification results obtained by two different sEMG datasets denote highrobustness of the novel algorithm. Moreover, further evaluation of the experimentalresults denotes that both the mis-trigger issue as well as power consumption are reducedwhen applying this novel algorithm.Additionally, to further utilize the potential of sEMG signals, another novel algorithmwhich aims at frequency domain features is presented in this thesis. This algorithmsuccessfully extracts the frequency domain features by applying singular system analysis(SSA) to the raw sEMG signals and converts them into frequency density maps (FDM).With the help of a proposed SNN, the algorithm is proved to be efficient for differenthand gestures. In addition, further evaluation of this algorithm indicates that itdemonstrates a significant reduction in computational costs, training time, powerconsumption whilst, at the same time, results in lower classification errors/mis-triggerswhen compared to other state of the art hand gesture recognition methodologies.

Degree

thesis:*
Name dc:type.qualificationname
phd
Level dc:type.qualificationlevel
doctoral-pg
Grantor dc:publisher.institution
University of Strathclyde
Year dc:date.issued
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Ke, Weijie
Advisors dc:contributor.advisor
  • Petropoulakis, L. (Lykourgos)
  • Soragahan, J. J.
Author Identifier dc:creator.authoridentifier
  • 201777615

Rights

Identifiers

dc:identifier.*
Identifier
T17648

Chain of custody

source
Harvested from
University of Strathclyde
Base URL
stax.strath.ac.uk/catalog/oai
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Ke, Weijie. Deep learning and sEMG signals based human hand gesture recognition. doctoral-pg thesis, University of Strathclyde, 2026. https://stax.strath.ac.uk/concern/theses/s1784m34x