Masters Theses

Permanent URI for this collection

https://hdl.handle.net/10919/9291

Browse

Now showing 1 - 20 of 20

Analyzing the complexity of bat flight to inspire the design of flapping-flight drones
Tyler, Adam Anthony Murphrey (Virginia Tech, 2024-08-22)
With their exceptionally maneuverable flapping flight, bats could serve as a model for enhancing the flight abilities for future drones. However, bat flight is extremely complex and there are many engineering restrictions that prevent a flapping-flight drone from replicating the many degrees of freedoms present in biology. Hence, to make design choices of which properties in a bats wing kinematics should be reproduced, the present research has evaluated two metrics from information and complexity theory to identify which regions of the bat flight apparatus are most complex and where coupling across features of the bat flight kinematics exists. The values were the complexity metric as a measure of variability and mutual information as a measure of coupling. Both measures were applied to ten experimentally obtained digital models of the flight kinematics in Ridley's leaf-nosed bats as well as the simulated kinematics of a flapping-flight drone inspired by the same bat type. The pilot results obtained indicate that both measures could be useful to discover which elements of flight kinematics should be looked into for understanding and reproducing the maneuvering flight of bats. However, a functional interpretation will require complementary, e.g., aerodynamic metrics.
Camera-based Recovery of Cardiovascular Signals from Unconstrained Face Videos Using an Attention Network
Deshpande, Yogesh Rajan (Virginia Tech, 2023-06-22)
This work addresses the problem of recovering the morphology of blood volume pulse (BVP) information from a video of a person's face. Video-based remote plethysmography methods have shown promising results in estimating vital signs such as heart rate and breathing rate. However, recovering the instantaneous pulse rate signals is still a challenge for the community. This is due to the fact that most of the previous methods concentrate on capturing the temporal average of the cardiovascular signals. In contrast, we present an approach in which BVP signals are extracted with a focus on the recovery of the signal morphology as a generalized form for the computation of physiological metrics. We also place emphasis on allowing natural movements by the subject. Furthermore, our system is capable of extracting individual BVP instances with sufficient signal detail to facilitate candidate re-identification. These improvements have resulted in part from the incorporation of a robust skin-detection module into the overall imaging-based photoplethysmography (iPPG) framework. We present extensive experimental results using the challenging UBFC-Phys dataset and the well-known COHFACE dataset. The source code is available at https://github.com/yogeshd21/CVPM-2023-iPPG-Paper.
Continuous Monitoring of High Risk Disaster Areas by Applying Change Detection to Free Satellite Imagery
Roush, Allison Granfield (Virginia Tech, 2024-06-11)
Natural disasters can happen anywhere causing damage to land and infrastructure. When these disasters occur in remote areas without much human traffic, it may take a long time for someone to notice that an event has occurred and to respond to it. Response time and damages could be reduced if the area could be remotely monitored. Many satellites pass over the Earth everyday collecting valuable imagery data that is free to access. However, this data can be difficult to process and use in practical applications such as monitoring an area for changes. Existing programs that use satellite imagery to monitor areas for changes can cost a significant amount of money making it inaccessible to most people. In this paper, a software program is introduced to automatically retrieve, process, and analyze free satellite imagery data and notify the user of significant changes in their area of interest (AOI). First, a software program was developed to automatically download a package of satellite imagery data from Planet Labs that met certain requirements for AOI, date, and cloud cover. A second software program was developed to download this data from the Google Cloud Storage (GCS) space and compare a current image to the composite of previous images in order to detect a change. This program then creates a figure to display the current image, the previous image, the difference area, and a summary table of the difference metrics. This figure is saved and emailed to the user if the differences are greater than the set threshold. This program is also capable of running automatically in the background of a computer every time it is logged in. The success of the program in correctly identifying areas of change was tested in three locations using historical satellite image data. The software was successful in identifying areas of change and delivering this information to the user in an easy to understand summary figure. Overall, the software was able to utilize free satellite imagery to detect changes in disaster areas and deliver a summary report to a user to take action showing that this software could be used in the future as an easy way to monitor disaster areas.
Data Centric Defenses for Privacy Attacks
Abhyankar, Nikhil Suhas (Virginia Tech, 2023-08-14)
Recent research shows that machine learning algorithms are highly susceptible to attacks trying to extract sensitive information about the data used in model training. These attacks called privacy attacks, exploit the model training process. Contemporary defense techniques make alterations to the training algorithm. Such defenses are computationally expensive, cause a noticeable privacy-utility tradeoff, and require control over the training process. This thesis presents a data-centric approach using data augmentations to mitigate privacy attacks. We present privacy-focused data augmentations to change the sensitive data submitted to the model trainer. Compared to traditional defenses, our method provides more control to the individual data owner to protect one's private data. The defense is model-agnostic and does not require the data owner to have any sort of control over the model training. Privacypreserving augmentations are implemented for two attacks namely membership inference and model inversion using two distinct techniques. While the proposed augmentations offer a better privacy-utility tradeoff on CIFAR-10 for membership inference, they reduce the reconstruction rate to ≤ 1% while reducing the classification accuracy by only 2% against model inversion attacks. This is the first attempt to defend model inversion and membership inference attacks using decentralized privacy protection.
Enhancing Road Safety through Machine Learning for Prediction of Unsafe Driving Behaviors
Sonth, Akash Prakash (Virginia Tech, 2023-08-21)
Road accidents pose a significant threat, leading to fatalities and injuries with far-reaching consequences. This study addresses two crucial challenges in road safety: analyzing traffic intersections to enhance safety by predicting potentially risky situations, and monitoring driver activity to prevent distracted driving accidents. Focusing on Virginia's intersections, we thoroughly examine traffic participant interactions to identify and mitigate conflicts, employing graph-based modeling of traffic scenarios to evaluate contributing parameters. Additionally, we leverage graph neural networks to detect and track potential crash situations from intersection videos, offering practical recommendations to enhance intersection safety. To understand the causes of risky behavior, we specifically investigate accidents resulting from distracted driving, which has become more prevalent due to advanced driver assistance systems in semi-autonomous vehicles. For monitoring driver activity inside vehicles, we propose the use of Video Transformers on challenging secondary driver activity datasets, incorporating grayscale and low-quality data to overcome limitations in capturing overall image context. Finally, we validate our predictions by studying attention modules and introducing explainability into the computer vision model. This research contributes to improving road safety by providing comprehensive analysis and recommendations for intersection safety enhancement and prevention of distracted driving accidents.
Estimation of Global Illumination using Cycle-Consistent Adversarial Networks
Oh, Junho (Virginia Tech, 2023-12-20)
The field of computer graphics has made significant progress over the years, transforming from simple, pixelated images to highly realistic visuals used across various industries including entertainment, fashion, and video gaming. However, the traditional process of rendering images remains complex and time-consuming, requiring a deep understanding of geometry, materials, and textures. This thesis introduces a simpler approach through a machine learning model, specifically using Cycle-Consistent Adversarial Networks (CycleGAN), to generate realistic images and estimate global illumination in real-time, significantly reducing the need for extensive expertise and time investment. Our experiments on the Blender and Portal datasets demonstrate the model's ability to efficiently generate high-quality, globally illuminated scenes, while a comparative study with the Pix2Pix model highlights our approach's strengths in preserving fine visual details. Despite these advancements, we acknowledge the limitations posed by hardware constraints and dataset diversity, pointing towards areas for future improvement and exploration. This work aims to simplify the complex world of computer graphics, making it more accessible and user-friendly, while maintaining high standards of visual realism.
General-Purpose Task Guidance from Natural Language in Augmented Reality using Vision-Language Models
Stover, Daniel James (Virginia Tech, 2024-06-12)
Augmented reality task guidance systems provide assistance for procedural tasks, which require a sequence of physical actions, by rendering virtual guidance visuals within the real-world environment. An example of such a task would be to secure two wood parts together, which could display guidance visuals indicating the user to pick up a drill and drill each screw. Current AR task guidance systems are limited in that they require AR system experts for use, require CAD models of real-world objects, or only function for limited types of tasks or environments. We propose a general-purpose AR task guidance approach and proof-of-concept system to generate guidance for tasks defined by natural language. Our approach allows an operator to take pictures of relevant objects and write task instructions for an end user, which are used by the system to determine where to place guidance visuals. Then, an end user can receive and follow guidance even if objects change location or environment. Guidance includes reusable visuals that display generic actions, such as our system's 3D hand animations. Our approach utilizes current vision-language machine learning models for text and image semantic understanding and object localization. We built a proof-of-concept system using our approach and tested its accuracy and usability in a user study. We found that all operators were able to generate clear guidance for tasks in an office room, and end users were able to follow the guidance visuals to complete the expected action 85.7% of the time without knowledge of their tasks. Participants rated that our system was easy to use to generate guidance visuals they expected.
Geometric Deep Learning for Healthcare Applications
Karwande, Gaurang Ajit (Virginia Tech, 2023-06-06)
This thesis explores the application of Graph Neural Networks (GNNs), a subset of Geometric Deep Learning methods, for medical image analysis and causal structure learning. Tracking the progression of pathologies in chest radiography poses several challenges in anatomical motion estimation and image registration as this task requires spatially aligning the sequential X-rays and modelling temporal dynamics in change detection. The first part of this thesis proposes a novel approach for change detection in sequential Chest X-ray (CXR) scans using GNNs. The proposed model CheXRelNet utilizes local and global information in CXRs by incorporating intra-image and inter-image anatomical information and showcases an increased downstream performance for predicting the change direction for a pair of CXRs. The second part of the thesis focuses on using GNNs for causal structure learning. The proposed method introduces the concept of intervention on graphs and attempts to relate belief propagation in Bayesian Networks (BN) to message passing in GNNs. Specifically, the proposed method leverages the downstream prediction accuracy of a GNN-based model to infer the correctness of Directed Acyclic Graph (DAG) structures given observational data. Our experimental results do not reveal any correlation between the downstream prediction accuracy of GNNs and structural correctness and hence indicate the harms of directly relating message passing in GNNs to belief propagation in BNs. Overall, this thesis demonstrates the potential of GNNs in medical image analysis and highlights the challenges and limitations of applying GNNs to causal structure learning.
H2OGAN: A Deep Learning Approach for Detecting and Generating Cyber-Physical Anomalies
Lin, Yen-Cheng (Virginia Tech, 2024-05-17)
The integration of Artificial Intelligence (AI) into water supply systems (WSSs) has revolutionized real-time monitoring, automated operational control, and predictive decision-making analytics. However, AI also introduces security vulnerabilities, such as data poisoning. In this context, data poisoning could involve the malicious manipulation of critical data, including water quality parameters, flow rates, and chemical composition levels. The consequences of such threats are significant, potentially jeopardizing public safety and health due to decisions being made based on poisoned data. This thesis aims to exploit these vulnerabilities in data-driven applications within WSSs. Proposing Water Generative Adversarial Networks, H2OGAN, a time-series GAN-based model designed to synthesize water data. H2OGAN produces water data based on the characteristics within the expected constraints of water data cardinality. This generative model serves multiple purposes, including data augmentation, anomaly detection, risk assessment, cost-effectiveness, predictive model optimization, and understanding complex patterns within water systems. Experiments are conducted in AI and Cyber for Water and Agriculture (ACWA) Lab, a cyber-physical water testbed that generates datasets replicating both operational and adversarial scenarios in WSSs. Identifying adversarial scenarios is particularly importance due to their potential to compromise water security. The datasets consist of 10 physical incidents, including normal conditions, sensor anomalies, and malicious attacks. A recurrent neural network (RNN) model, i.e., gated recurrent unit (GRU), is used to classify and capture the temporal dynamics those events. Subsequently, experiments with real-world data from Alexandria Renew Enterprises (AlexRenew), a wastewater treatment plant in Alexandria, Virginia, are conducted to assess the effectiveness of H2OGAN in real-world applications.
Harnessing the Power of Self-Training for Gaze Point Estimation in Dual Camera Transportation Datasets
Bhagat, Hirva Alpesh (Virginia Tech, 2023-06-14)
This thesis proposes a novel approach for efficiently estimating gaze points in dual camera transportation datasets. Traditional methods for gaze point estimation are dependent on large amounts of labeled data, which can be both expensive and time-consuming to collect. Additionally, alignment and calibration of the two camera views present significant challenges. To overcome these limitations, this thesis investigates the use of self-learning techniques such as semi-supervised learning and self-training, which can reduce the need for labeled data while maintaining high accuracy. The proposed method is evaluated on the DGAZE dataset and achieves a 57.2\% improvement in performance compared to the previous methods. This approach can prove to be a valuable tool for studying visual attention in transportation research, leading to more cost-effective and efficient research in this field.
Hierarchy Aligned Commonality Through Prototypical Networks: Discovering Evolutionary Traits over Tree-of-Life
Manogaran, Harish Babu (Virginia Tech, 2024-10-11)
A grand challenge in biology is to discover evolutionary traits, which are features of organisms common to a group of species with a shared ancestor in the Tree of Life (also referred to as phylogenetic tree). With the recent availability of large-scale image repositories in biology and advances in the field of explainable machine learning (ML) such as ProtoPNet and other prototype-based methods, there is a tremendous opportunity to discover evolutionary traits directly from images in the form of a hierarchy of prototypes learned at internal nodes of the phylogenetic tree. However, current prototype-based methods are mostly designed to operate over a flat structure of classes and face several challenges in discovering hierarchical prototypes on a tree, including the problem of learning over-specific features at internal nodes in the tree. To overcome these challenges, we introduce the framework of Hierarchy aligned Commonality through Prototypical Networks (HComP-Net), which learns common features shared by all descendant species of an internal node and avoids the learning of over-specific prototypes. We empirically show that HComP-Net learns prototypes that are of high accuracy, semantically consistent, and generalizable to unseen species in comparison to baselines. While we focus on the biological problem of discovering evolutionary traits, our work can be applied to any domain involving a hierarchy of classes.
Improvement in Frame Prediction using Optical Flow
Wormack Jr, Craig Frederick Luther (Virginia Tech, 2023-07-06)
Future frame prediction is a difficult but useful problem to solve in deep learning. The technology can be used to predict future occurrences in a video, anticipate anomalies, and aid autonomous devices in smart decision making. Although there is potential with frame prediction technology, there is still progress that needs to be made with it. As the predicted frame becomes farther away from the last input frame, the image becomes blurry and distorted. This indicates that the model is more uncertain about the motion occurring in the image frame. To reduce model uncertainty shown in predictions, optical flow information from each video was extracted and combined with the video frames. An optical flow-based approach is uncommon in frame prediction and has not been implemented with a fully Convolutional Neural Network (CNN) based architecture. In this work, the change in image quality evaluation metrics and overall image quality is analyzed across 4 different datasets between a state-of-the-art frame prediction model and a modified model that combines optical flow information. The results demonstrate that adding optical flow information improves the model Mean Squared Error (MSE) by 4.11% and its Structural Similarity Index Metric (SSIM) by 0.41% for the Moving MNIST dataset. Optical flow improved the SSIM value of Taxi BJ, KTH, and KITTI by 0.02%, 0.011%, and 1.297% respectively. While there was a consistent improvement in performance, the models still need more improvement in terms of the quality of images predicted in the distant future.
Intelligently Leveraging Multi-Channel Image Processing Neural Networks for Multi-View Co-Channel Signal Detection
Koppikar, Nidhi Nitin (Virginia Tech, 2024-08-19)
The evolution of technology and gadgets has led to a significant increase in the number of transmitted signals, making RF sensing more complex than ever. Challenges such as signal interference and the lack of prior information about all signal parameters further complicate the task. To address this challenge, researchers have explored machine learning and deep learning approaches to generalize solutions for real-world sensing problems. In this thesis, we focus on two key issues in RF signal detection using deep learning. Firstly, we tackle the problem of increasing signal detection coverage by utilizing multiple resolution eigengram images derived from a bank of channelizers. These channelizers, varying in size, are adept at sensing different types of signals, such as low duration or low bandwidth signals. Channelizer deconfliction is a known challenge in RFML. We use YOLO, a deep learning algorithm, to deconflict the outputs from different channelizers to avoid overreporting. YOLO's ability to handle three channels makes it ideal for our study as we also use three channelizers. While our approach is not dependent on YOLO, it provides a good testing ground for this study. To address signal overlap, we utilize an eigengram image capturing the overlap region between signals. By overlaying this eigengram onto the original, we create a composite image highlighting the overlap. We train another YOLO model using two channels, one for each eigengram, enabling detection even with over 50 percent overlap. This work is versatile and promising, extending to other signal visualizations. It has significant potential for wireless industry applications and sets the stage for further RFML research.
A Machine Learning Approach to Recognize Environmental Features Associated with Social Factors
Diaz-Ramos, Jonathan (Virginia Tech, 2024-06-11)
In this thesis we aim to supplement the Climate and Economic Justice Screening Tool (CE JST), which assists federal agencies in identifying disadvantaged census tracts, by extracting five environmental features from Google Street View (GSV) images. The five environmental features are garbage bags, greenery, and three distinct road damage types (longitudinal, transverse, and alligator cracks), which were identified using image classification, object detection, and image segmentation. We evaluate three cities using this developed feature space in order to distinguish between disadvantaged and non-disadvantaged census tracts. The results of the analysis reveal the significance of the feature space and demonstrate the time efficiency, detail, and cost-effectiveness of the proposed methodology.
Machine Learning Classification of Gas Chromatography Data
Clark, Evan Peter (Virginia Tech, 2023-08-28)
Gas Chromatography (GC) is a technique for separating volatile compounds by relying on adherence differences in the chemical components of the compound. As conditions within the GC are changed, components of the mixture elute at different times. Sensors measure the elution and produce data which becomes chromatograms. By analyzing the chromatogram, the presence and quantity of the mixture's constituent components can be determined. Machine Learning (ML) is a field consisting of techniques by which machines can independently analyze data to derive their own procedures for processing it. Additionally, there are techniques for enhancing the performance of ML algorithms. Feature Selection is a technique for improving performance by using a specific subset of the data. Feature Engineering is a technique to transform the data to make processing more effective. Data Fusion is a technique which combines multiple sources of data so as to produce more useful data. This thesis applies machine learning algorithms to chromatograms. Five common machine learning algorithms are analyzed and compared, including K-Nearest Neighbour (KNN), Support Vector Machines (SVM), Convolutional Neural Network (CNN), Decision Tree, and Random Forest (RF). Feature Selection is tested by applying window sweeps with the KNN algorithm. Feature Engineering is applied via the Principal Component Analysis (PCA) algorithm. Data Fusion is also tested. It was found that KNN and RF performed best overall. Feature Selection was very beneficial overall. PCA was helpful for some algorithms, but less so for others. Data Fusion was moderately beneficial.
Methodology for Zero-Cost Auto-tuning of Embedded PID Controllers for Actuators: A Study on Proportional Valves in Micro Gas Chromatography Systems
Korada, Divya Tarana (Virginia Tech, 2024-06-21)
This thesis describes the implementation of zero-cost auto-tuning techniques for embedded Proportional Integral and Derivative (PID) controllers, specifically focusing on their application in the control of proportional valves within Micro Gas Chromatography (uGC) systems. uGC systems are miniaturized versions of conventional GC systems, and require precise temperature, flow and pressure control for the micro-fabricated preconcentrators and micro columns. PID controllers are widely used in process control applications due to their simplicity and effectiveness. The Commercial Off The Shelf (COTS) available controllers are expensive, bulky, need system compatibility and have high lead times. The proposed auto-tuner features simple Python-implemented empirical calculations based on Ziegler Nichols relay-based PID tuning method to determine the optimal PID gains. Leveraging Wi-Fi the system enables tuning for any embedded platform while visualizing transient response through the Graphical User Interface (GUI). The embedded-GUI interface provides a customizable auto-tuning experience extending usage across diverse temperature, pressure and flow regulation applications in environmental analysis. Specifically for uGC systems, the GUI integrates with existing hardware stack using minor software enhancements to enable rapid, automated PID tuning for thermal and flow control applications. The performance is analyzed by evaluating response metrics including overshoot, rise time, and steady-state error.
Non-invasive Estimation of Skin Chromophores Using Hyperspectral Imaging
Karambor Chakravarty, Sriya (Virginia Tech, 2024-03-07)
Melanomas account for more than 1.7% of global cancer diagnoses and about 1% of all skin cancer diagnoses in the United States. This type of cancer occurs in the melanin-producing cells in the epidermis and exhibits distinctive variations in melanin and blood concentration values in the form of skin lesions. The current approach for evaluating skin cancer lesions involves visual inspection with a dermatoscope, typically followed by biopsy and histopathological analysis. However, to decrease the risk of misdiagnosis in this process requires invasive biopsies, contributing to the emotional and financial distress of patients. The implementation of a non-invasive imaging technique to aid the analysis of skin lesions in the early stages can potentially mitigate these consequences. Hyperspectral imaging (HSI) has shown promise as a non-invasive technique to analyze skin lesions. Images taken of human skin using a hyperspectral camera are a result of numerous elements in the skin. Being a turbid, inhomogeneous material, the skin has chromophores and scattering agents, which interact with light and produce characteristic back-scattered energy that can be harnessed and examined with an HSI camera. To achieve this in this study, a mathematical model of the skin is used to extract meaningful information from the hyperspectral data in the form of parameters such as melanin concentration, blood volume fraction and blood oxygen saturation in the skin. The human skin is modelled as a bi-layer planar system, whose surface reflectance is theoretically calculated using the Kubelka-Munk theory and absorption laws by Beer and Lambert. The model is evaluated for its sensitivity to the parameters and then fitted to measured hyperspectral data of four volunteer subjects in different conditions. Mean values of melanin, blood volume fraction and oxygen saturation obtained for each of the subjects are reported and compared with theoretical values from literature. Sensitivity analysis revealed wavelengths and wavelength groups which resulted in maximum change in percentage reflectance calculated from the model were 450 and 660 nm for melanin, 500 - 520 nm and 590 - 625 nm for blood volume fraction and 606, 646 and 750 nm for blood oxygen saturation.
Non-invasive estimation of skin chromophores using Hyperspectral Imaging
Karambor Chakravarty, Sriya (Virginia Tech, 2023-08-21)
Melanomas account for more than 1.7% of global cancer diagnoses and about 1% of all skin cancer diagnoses in the United States. This type of cancer occurs in the melanin-producing cells in the epidermis and exhibits distinctive variations in melanin and blood concentration values in the form of skin lesions. The current approach for evaluating skin cancer lesions involves visual inspection with a dermatoscope, typically followed by biopsy and histopathological analysis. However, this process, to decrease the risk of misdiagnosis, results in unnecessary biopsies, contributing to the emotional and financial distress of patients. The implementation of a non-invasive imaging technique to aid the analysis of skin lesions in the early stages can potentially mitigate these consequences. Hyperspectral imaging (HSI) has shown promise as a non-invasive technique to analyze skin lesions. Images taken of human skin using a hyperspectral camera are a result of numerous elements in the skin. Being a turbid, inhomogeneous material, the skin has chromophores and scattering agents, which interact with light and produce characteristic back-scattered energy that can be harnessed and examined with an HSI camera. In this study, a mathematical model of the skin is used to extract meaningful information from the hyperspectral data in the form of melanin concentration, blood volume fraction and blood oxygen saturation in the skin. The human skin is modelled as a bi-layer planar system, whose surface reflectance is theoretically calculated using the Kubelka-Munk theory and absorption laws by Beer and Lambert. Hyperspectral images of the dorsal portion of three volunteer subjects' hands 400 - 1000 nm range, were used to estimate the contributing parameters. The mean and standard deviation of these estimates are reported compared with theoretical values from the literature. The model is also evaluated for its sensitivity with respect to these parameters, and then fitted to measured hyperspectral data of three volunteer subjects in different conditions. The wavelengths and wavelength groups which were identified to result in the maximum change in percentage reflectance calculated from the model were 450 and 660 nm for melanin, 500 - 520 nm and 590 - 625 nm for blood volume fraction and 606, 646 and 750 nm for blood oxygen saturation.
PID Auto-Tuning and Control System for Heaters in μGC Systems
Gupta, Poonam (Virginia Tech, 2023-03-31)
Micro gas chromatography (μGC) system is a miniaturized and portable version of the conventional GC system, suitable for various applications such as healthcare and environmental analysis. The process of gas chromatography requires precise temperature control for the micro-fabricated preconcentrators and separation columns used since temperature changes directly affect retention time. Proportional Integral and Derivative (PID) controllers provide reliable temperature control and can be tuned to obtain the desired response. The conventional method of tuning the PID control parameters by trial and error is a tedious process and time-consuming process. This thesis aims to develop a PID auto-tuning and control system for auto-tuning microfabricated heaters in modular μGC systems. The developed system is based on the Ziegler Nichols rule-based PID tuning method for closed-loop systems, which uses the relay response of the micro-heater to calculate the PID tuning parameters. The system also includes an analysis system to verify the performance of the PID-tuned values and a tuning system where the PID values can be further tuned to obtain more precise control for the heaters. The aim of developing this system is to reduce the effective tuning time for heaters while satisfying the control requirements. In this thesis, we discuss the tuning methodology and the implementation of the PID tuning and control system, followed by a performance evaluation of the heaters tuned using the proposed system is discussed.
A Temporal Encoder-Decoder Approach to Extracting Blood Volume Pulse Signal Morphology from Face Videos
Li, Fulan (Virginia Tech, 2023-07-05)
This thesis considers methods for extracting blood volume pulse (BVP) representations from video of the human face. Whereas most previous systems have been concerned with estimating vital signs such as average heart rate, this thesis addresses the more difficult problem of recovering BVP signal morphology. We present a new approach that is inspired by temporal encoder-decoder architectures that have been used for audio signal separation. As input, this system accepts a temporal sequence of RGB (red, green, blue) values that have been spatially averaged over a small portion of the face. The output of the system is a temporal sequence that approximates a BVP signal. In order to reduce noise in the recovered signal, a separate processing step extracts individual pulses and performs normalization and outlier removal. After these steps, individual pulse shapes have been extracted that are sufficiently distinct to support biometric authentication. Our findings demonstrate the effectiveness of our approach in extracting BVP signal morphology from facial videos, which presents exciting opportunities for further research in this area. The source code is available at https://github.com/Adleof/CVPM-2023-Temporal-Encoder-Decoder-iPPG

Browse

Browsing Masters Theses by Author "Abbott, Amos L."

Results Per Page

Sort Options