<?xml version="1.0" encoding="UTF-8"?><?xml-stylesheet type="text/xsl" href="static/style.xsl"?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2026-11-22T21:03:49Z</responseDate><request verb="GetRecord" identifier="oai:uwspace.uwaterloo.ca:10012/19252" metadataPrefix="dim">https://dspacemainprd01.lib.uwaterloo.ca/server/oai/request</request><GetRecord><record><header><identifier>oai:uwspace.uwaterloo.ca:10012/19252</identifier><datestamp>2025-03-26T23:35:14Z</datestamp><setSpec>com_10012_9900</setSpec><setSpec>com_10012_1</setSpec><setSpec>col_10012_9908</setSpec><setSpec>col_10012_6</setSpec></header><metadata><dim:dim xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:doc="http://www.lyncode.com/xoai" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.dspace.org/xmlns/dspace/dim http://www.dspace.org/schema/dim.xsd">
   <dim:field mdschema="dc" element="contributor" qualifier="advisor">Karray, Fakhri</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="author" authority="a5cbf5aa-a051-4c6c-85f5-8ae7a0db772a" confidence="600">Nooruddin, Sheikh</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="accessioned">2023-04-05T16:09:43Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="available">2024-04-05T04:50:04Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="issued">2023-04-05</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="submitted">2023-03-27</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="uri">http://hdl.handle.net/10012/19252</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="abstract" lang="en">Human Activity Recognition (HAR) is the process of automatic recognition of Activities of&#xd;
Daily Life (ADL) from human motion data captured in various data modalities by wearable and&#xd;
ambient sensors. Advances in deep learning, especially Convolutional Neural Networks (CNN)&#xd;
have revolutionized intelligent frameworks such as HAR systems by effectively and efficiently&#xd;
inferring human activity from various modalities of data. However, the training and inference&#xd;
of CNNs are often resource-intensive. Recent research developments are focused on bringing&#xd;
the effectiveness of CNNs in resource constrained edge devices through Tiny Machine Learning&#xd;
(TinyML). TinyML aims to optimize these models in terms of compute and memory requirements&#xd;
- aiming to make them suitable for always-on resource constrained devices - leading to a reduc-&#xd;
tion in communication latency and network traffic for HAR frameworks. In this thesis, at first,&#xd;
we provide a benchmark to understand these trade-offs among variations of CNN network archi-&#xd;
tectures, different training methodologies, and different modalities of data in the context of HAR,&#xd;
TinyML, and edge devices. We tested and reported the performance of CNN and Depthwise Sep-&#xd;
arable Convolutional Neural Network (DSCNN) models as well as two training methodologies:&#xd;
Quantization Aware Training (QAT) and Post Training Quantization (PTQ) on five commonly&#xd;
used benchmark datasets containing image and time-series data: UP-Fall, Fall Detection Dataset&#xd;
(FDD), PAMAP2, UCI-HAR, and WISDM. We also deployed and tested the performance of the&#xd;
model-based standalone applications on multiple commonly available resource constrained edge&#xd;
devices in terms of inference time and power consumption. Later, we focus on HAR from video&#xd;
data sources. We proposed a two-stream multi-resolution fusion architecture for HAR from video&#xd;
data modality. The context stream takes a resized image as input and the fovea stream takes the&#xd;
cropped center portion of the resized image as input, reducing the overall dimensionality. Due&#xd;
to camera bias, objects of interest are often situated in the center of the frame. We tested two&#xd;
quantization methods: PTQ and QAT to optimize these models for deployment in edge devices&#xd;
and tested the performance in two challenging video datasets: KTH and UCF11. We performed&#xd;
ablation studies to validate the two-stream model performance. We deployed the proposed ar-&#xd;
chitecture in commercial resource constrained devices and monitored their performance in terms&#xd;
of inference latency and power consumption. The results indicate that the proposed architecture&#xd;
clearly outperforms other relevant single-stream models tested in this work in terms of accuracy,&#xd;
precision, recall, and F1 score while also reducing the overall model size. The experimental&#xd;
results in this thesis demonstrate the effectiveness and feasibility of TinyML for HAR from mul-&#xd;
timodal data sources in edge devices.</dim:field>
   <dim:field mdschema="dc" element="language" qualifier="iso" lang="en">en</dim:field>
   <dim:field mdschema="dc" element="publisher" lang="en">University of Waterloo</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Tiny Machine Learning (TinyML)</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Human Activity Recognition (HAR)</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Deep Learning (DL)</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Computer Vision (CV)</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en">Image Processing</dim:field>
   <dim:field mdschema="dc" element="title" lang="en">On the Design of Efficient Deep Learning Methods for Human Activity Recognition in Resource Constrained Devices</dim:field>
   <dim:field mdschema="dc" element="type" lang="en">Master Thesis</dim:field>
   <dim:field mdschema="dc" element="pending">false</dim:field>
   <dim:field mdschema="uws-etd" element="degree" qualifier="department" lang="en">Electrical and Computer Engineering</dim:field>
   <dim:field mdschema="uws-etd" element="degree" qualifier="discipline" lang="en">Electrical and Computer Engineering</dim:field>
   <dim:field mdschema="uws-etd" element="degree" qualifier="grantor" lang="en">University of Waterloo</dim:field>
   <dim:field mdschema="uws-etd" element="degree" lang="en">Master of Applied Science</dim:field>
   <dim:field mdschema="uws-etd" element="embargo" qualifier="terms" lang="en">1 year</dim:field>
   <dim:field mdschema="uws" element="contributor" qualifier="advisor">Karray, Fakhri</dim:field>
   <dim:field mdschema="uws" element="contributor" qualifier="affiliation1" lang="en">Faculty of Engineering</dim:field>
   <dim:field mdschema="uws" element="published" qualifier="city" lang="en">Waterloo</dim:field>
   <dim:field mdschema="uws" element="published" qualifier="country" lang="en">Canada</dim:field>
   <dim:field mdschema="uws" element="published" qualifier="province" lang="en">Ontario</dim:field>
   <dim:field mdschema="uws" element="typeOfResource" lang="en">Text</dim:field>
   <dim:field mdschema="uws" element="peerReviewStatus" lang="en">Unreviewed</dim:field>
   <dim:field mdschema="uws" element="scholarLevel" lang="en">Graduate</dim:field>
   <dim:field mdschema="others" element="access-status">open.access</dim:field>
</dim:dim>
</metadata></record></GetRecord></OAI-PMH>