Doctor of Philosophy, University of Toledo, 2019, Engineering (Computer Science)

Multi-modal interaction is a type of Human-Computer Interaction (HCI), which involves a combination of sensory and representation modalities. In human-to-human interactions, the participants usually make use of all modalities and media available. These include speech, gestures, facial expressions, eye movements, and documents. These modalities can be captured using different types of sensors such as a microphone for voice, a camera or live video for gesture recognition, and a touchscreen for touch. The redundancies often introduced in this activity are one of the ways in which humans ensure messages are understood without the use of any technology. Multimodal interaction plays an important role in resolving such ambiguities. Their prowess to emanate unambiguous information exchange between the two collaborators make these systems more reliable, efficient, less error prone and capable of putting up complex and varied situation and tasks. Emotion recognition is a realm of HCI that follow the multimodal aspect to achieve more accurate and more natural results. Prodigious uses of affective identification in e-learning, marketing, security, health sciences etc. has resulted in the increase in demand of high precision emotion recognition systems. Machine learning is getting its feet wet to ameliorate the process by tweaking the architectures or by wielding high quality databases. This dissertation presents an insight of the work done in the areas of multi-modal HCI and their use for emotion recognition. Fusion, an important component in the architecture of the multi-modal HCI forms the cornerstone of this research work. Its implementation in various forms is discussed and implemented. Phase I of the research begins with a proposition for the fusion of two modalities at the grammar level while preserving the semantic of the modalities. This is achieved by inferring grammars and then combining those grammars using operators of the GA. The related results, assumptions (open full item for complete abstract)

Doctor of Philosophy (PhD), Wright State University, 2014, Computer Science and Engineering PhD

To construct a complete representation of a scene with environmental obstacles such as fog, smoke, darkness, or textural homogeneity, multisensor video streams captured in diferent modalities are considered. A computational method for automatically fusing multimodal image streams into a highly informative and unified stream is proposed. The method consists of the following steps: 1. Image registration is performed to align video frames in the visible band over time, adapting to the nonplanarity of the scene by automatically subdividing the image domain into regions approximating planar patches 2. Wavelet coefficients are computed for each of the input frames in each modality 3. Corresponding regions and points are compared using spatial and temporal information across various scales 4. Decision rules based on the results of multimodal image analysis are used to combine the wavelet coefficients from different modalities 5. The combined wavelet coefficients are inverted to produce an output frame containing useful information gathered from the available modalities Experiments show that the proposed system is capable of producing fused output containing the characteristics of color visible-spectrum imagery while adding information exclusive to infrared imagery, with attractive visual and informational properties.

Master of Music (MM), Ohio University, 2024, Music Therapy (Fine Arts)

This thesis investigated the role of musical fusion in music therapy, emphasizing its impact on therapeutic relationships and cultural responsiveness. Using a narrative inquiry approach, the study explored the experiences of board-certified music therapists who integrate diverse musical styles into their practice. Despite a small sample size, thematic analysis revealed that musical fusion can enhance therapeutic relationships more quickly than non-musical fusion methods. The findings underscored the importance of cultural responsiveness and suggested that continuous training in diverse musical genres is essential for music therapists. Recommendations for future research include examining the long-term effects of musical fusion and comparing its effectiveness with non-musical fusion techniques. This study provided valuable insights into the benefits of incorporating cultural elements in music therapy.

Doctor of Philosophy, Case Western Reserve University, 2021, Molecular Virology

Extracellular vesicles (EVs) are lipid-bilayer enclosed, cell-derived nanoparticles released constitutively by most cells. EVs carry proteins and nucleic acids from their cells of origin, facilitating intercellular communication through a variety of different pathways. These include signaling through cellular receptors, triggering immune responses, and delivering bioactive cargos after EV internalization. Viruses share many characteristics with EVs, including size, structure, and biogenesis pathways. Virally infected cells continue to release EVs, resulting in the production of virus-induced EVs that carry viral elements like proteins or RNAs. Viral fusion proteins – viral envelope glycoproteins that mediate virus/cell membrane fusion – are often incorporated by virus-induced EVs. In this work, we investigated the effect that viral fusion protein incorporation has on EV communication and functionality. First, we investigated the spike protein of the current pandemic virus SARS-CoV-2. Interestingly, we found that EVs released from spike expressing cells did incorporate this type I viral fusion protein. These spike(+) EVs displayed the spike protein in a conformation available for antibody binding, allowing the EVs to serve as decoy targets for neutralizing antibodies. This held true with both convalescent patient serum and a commercial neutralizing antibody, with the effect of reducing virus neutralization and increasing infection. The second component of this work concerned the development of EV/cell membrane fusion assays to monitor EV cargo delivery. This EV communication pathway is particularly important in disease, where EVs are thought to be capable of influencing cells through membrane fusion and delivery of bioactive cargos to cell cytosol. We successfully developed two EV fusion assays, based on delivery of the proteins beta-lactamase and Cre-recombinase. In the third and final component of this work, we utilized these fusion assays to investigate the in (open full item for complete abstract)

Master of Science in Engineering, Youngstown State University, 2020, Department of Mechanical, Industrial and Manufacturing Engineering

Laser Powder Bed Fusion (LBPF) allows for unparalleled freedom in design to manufacture complicated structures in high performance materials. Due to the advancement of additive manufacturing technologies, 3D printed parts have moved from the R&D phase to the development phase, with the aerospace industry having adapted this technology to cater small batch replacement parts mainly for an aging fleet of aircraft. The goal of this research was to systematically generate defects, mainly lack-of-fusion defects, to understand the mechanical and corrosion behavior of these defects in parts that are susceptible to flight criticality and safety criticality. This work investigated the influence of the main Selective Laser Melting process parameters (laser power, travel velocity, hatch spacing, and layer thickness) on the defect characteristics using an AlSi10Mg alloy. Five studies were conducted to analyze, evaluate, and compare the defect nature, including linearity, in different orientations and builds. Seventy-seven sample coupons were manufactured from two parameter development builds and image analysis was performed using Image J and Photoshop. Optical microscopy and X-ray CT imaging were the methods used for defect detection. The results showed that for decreasing energy density, the defect density and defect size increase which results in the decrease of the % average relative density, for the set of process parameters investigated, with lack-of-fusion defects predominantly forming at energy densities below 35 J/mm3. Following defect characterization, the effects of each of the four major process parameters were interpreted using a DOE approach with the help of regression and ANOVA testing. Hatch spacing proved to be the most significant process parameter affecting the defect density, while the layer thickness showed the most significant effect when predicting the average defect dimensions and the ratio of defect length to height for the set of process parameters (open full item for complete abstract)

Master of Science in Engineering, Youngstown State University, 2019, Department of Mechanical, Industrial and Manufacturing Engineering

Additive manufacturing is one of the more recent advances in manufacturing technology. Additive manufacturing processes allow for the creation of parts in a layer-by-layer fashion. There are several materials that can be used in additive manufacturing processes including metal, ceramic, and polymers which each presenting their own challenges. This work focuses on metal based additive manufacturing parts made out of AlSi10Mg using a process called laser powder bed fusion. Laser powder bed fusion is one of the three major metal additive manufacturing processes with the other two being multi-pass welding and direct energy deposition. One of many challenges that occur with the laser power bed fusion process is minimizing the residual stresses and distortion that are present in the part during and after the build. During the early days of additive manufacturing that was mostly done through a trial-and-error process where multiple version of a part would be printed until a desired outcome was achieved, and this was often very expensive, and time consuming. There has been plenty of research in developing simulation models in order to predict the distortions and stresses that developed during the additive manufacturing process. These simulations allowed engineers to optimize parts before they were printed, and thus reduce the number of wasted prints. This work demonstrates and validates use of a software package call Autodesk Netfabb Simulation in order to find the optimal orientation of a complex part. The optimal orientation was selected for three categories: distortion, stress, and printability. Optimal orientations were selected from a selection of 23 orientations that were simulated. To validate the simulations, two test parts along with three of the aforementioned orientations were printed and measured using 3D scanning while still the build plate. The result of this was that the optimal orientation was different for each of three criteria meaning it is up to the part (open full item for complete abstract)

Doctor of Philosophy (Ph.D.), University of Dayton, 2018, Engineering

Visual object tracking is an important research area within computer vision. Object tracking has applications in security, surveillance, robotics, and safety systems. In generic single object tracking, the problem is constrained to short-term tracking where the target is initialized using its location in a single frame and the tracker is not reinitialized. This is challenging because trackers must update the target model using predicted targets in later frames. However, this has a large potential to cause model drift as errors are introduced over time. Additional challenges that are present in visual tracking include illumination changes, partial and full occlusions, deformation of the target, viewpoints changes, scale change, complex backgrounds and clutter, and similar objects in the scene. A widely used strategy for improved tracking is to combine various complementary features. Combination strategies are varied in how they use the multiple features or trackers. Adaptive fusion is performed by basing the weighting on the value of individual estimates in previous frames. The proposed tracking scheme takes inspiration from human vision to reduce the risk of tracking errors. In our proposed tracking scheme, the learned adaptive feature fusion (LAFF) method, a robust modular tracker is created by adaptability updating the weighting scheme based on a trained system for scoring each estimator. This is accomplished by first researching previous feature fusion techniques and examining their shortcomings. A variance ratio based method for adaptive feature fusion (AFF) is developed and evaluated. Next, a machine learning based method is created to help further improve robustness for the tracker. The LAFF method is an extension of AFF that teaches a machine learned regressor to generate fusion weights for a set of features. A suite of diverse features is selected for fast and accurate tracking, while also demonstrating the advantage of adaptive fusion. These featu (open full item for complete abstract)

Master of Science, Miami University, 2017, Computational Science and Engineering

The tokamak fusion reactor is one of the most promising and well-developed designs for fusion energy production. Scientists around the world use tokamaks to research methods of generating electrical energy from the fusion reaction. Furthermore, efforts are now underway to design a new large sized tokamak, Chinese Fusion Engineering Test Reactor (CFETR), with the aim of demonstrating fusion energy as a viable source of power. As this project is just beginning, it is necessary to evaluate new and emerging technologies that can be used in this endeavor. Power electronics play a crucial role in fusion energy research and are the focus of the thesis. The High Power, Power Supplies (HPPS) transforms electrical energy from the grid into AC and DC signals with extremely high voltage and current magnitudes. For example, the Neutral Beam Injectors (NBI) require a power source that is capable of generating 110 kVDC at nearly 10 MW. This thesis evaluates two new types of technology for use in the NBI power supply; the utilization of new semiconductor switching devices and the application of new circuit topologies. The switching devices used in the existing HPPS all utilize silicon based semiconductors. Within the last ten years, new devices created from Wide Bandgap (WBG) semiconductors have become commercially available. This thesis demonstrates that gains in efficiency are possible by utilizing WBG based power devices in the NBI power supply. It also explores the application of a Modular Multilevel Converters (MMC) as a replacement to the existing topology used in the NBI power supply. It shows that the MMC can be used as both a rectifier for the NBI and an active filter for the electrical grid. The case study for this work is the NBI power supply of Experimental Advanced Superconducting Tokamak in Hefei, China. However, the findings are applicable to CFETR as well as other future designs of tokamak reactors.

Master of Science, University of Toledo, 2004, Bioengineering

In order to promote solid fusion across a decompressed spinal segment, inter-body spacers/cages are used with and without posterior instrumentation to provide an initial “rigid” fixation of the segment. Inter-body spacers (cages) of various shapes (e.g., rectangular, cylindrical) and materials are currently available on the market. Important factors affecting the biomechanics of the fused segment are (i) cage shape and placement, (ii) cage material property (iii) surgical approach used –posterior vs. antero lateral (iv) cage with additional instrumentation. The objective of this study is to address change in the stability and stress patterns associated with the various factors described above. A cadaveric study using established protocols and a finite element (FE) study were conducted. For the cadaveric study, nine fresh ligamentous lumbar spine specimens (L1-S2) were radiographed out of which six specimens were prepared for testing by fixing a base to the sacrum and a loading frame to the top-most vertebra. Each specimen was subjected to pure moment (6 Nm in steps of 1.5 Nm) in six loading modes: flexion, extension, right and left lateral bending, and right and left axial rotation. The load-displacement data was collected in a sequential manner for the following cases: 1) intact spine, 2) insertion of rectangular cages (Vertebral spacer PR, Synthes, Inc.), 3) fixation with posterior instrumentation, 4) fatiguing the instrumented spine. The relative motion of L4 with respect to L5 was calculated for all these cases. A validated three-dimensional, nonlinear FE model of lumbar spine from L3-L5 was used. The model was modified to simulate the bilateral placement of cages alone. Contact surfaces were defined between the cages and the endplates to simulate the bone-implant interface. The cages were placed using posterior approach and left antero lateral approach to see the effect of the surgical approach on the stability of the segment. In the FE model with cage placed u (open full item for complete abstract)

Master of Science, The Ohio State University, 2009, Electrical and Computer Engineering

This thesis addresses the problem of detecting and tracking objects in a scene, using a distributed set of sensing devices in different locations, and in general use a mix of different sensing modalities. The goal is to combine data in an efficient but statistically principled way to realize optimal or near-optimal detection and tracking performance. Using the Bayesian framework of measurement likelihood, sensor data can be combined in a rigorous manner to produce a concise summary of knowledge of a target's location in the state-space. This framework allows sensor data to be fused across time, space and sensor modality. When target motion and sensor measurements are modeled correctly, these “likelihood maps” are optimal combinations of sensor data. By combining all data without thresholding for detections, targets with low signal to noise ratio (SNR) can be detected where standard detection algorithms may fail. For estimating the location of multiple targets, the likelihood ratio is used to provide a sub-optimal but useful representation of knowledge of the state space. As the calculation cost of computing likelihood or likelihood ratio maps over the entire state space is prohibitively high for most practical applications, an approximation computed in a distributed fashion is proposed and analyzed. This distributed method is tested in simulation for multiple sensor modalities, displaying cases where it is and is not a good approximation of central calculation. Detection and tracking examples using measured data from multi-modal sensors (Radar, EO, Seismic) are also presented.

Doctor of Philosophy, Case Western Reserve University, 2011, Biomedical Engineering

PET imaging system delivers abundant functional information which is complementary to the anatomical information provided by CT images. The purpose of this research is to improve the physician's ability to localize and delineate the extent of the tumor by incorporation of the PET images into radiation therapy treatment planning. A machine-based CT-PET fiducial fusion method was implemented for head and neck carcinoma radiation therapy. In this method, the field arrangements are aligned relative to the fixed treatment machine isocenter and patients are imaged in actual treatment positions. A fiducial registration error (FRE) of 1 mm was found for this fiducial fusion method. The target registration error (TRE) of seven anatomical landmarks was measured to evaluate the accuracy of this method. The results were compared with a manual and a mutual information based automatic fusion method. Statistical analysis showed there was no significant difference of TREs between the fiducial fusion method and the manual method which is considered to be most accurate in this research. In addition, a new thresholding PET image segmentation method was proposed using a lookup table which consists of the recovered activity concentration ratios and the initial estimates of target volume. To validate the proposed segmentation method, a Jaszczak phantom containing hollow spheres with variable size and FDG concentration contrast ratio was scanned in different PET scanners. The average uncertainty of the volume estimation by the proposed method was 11.2% for spheres greater than 2.5 mL, which were comparable or superior to those determined by contrast-oriented method and iterative threshold method (ITM). This new segmentation method was also applied to the PET images of ten patients with solitary lung metastases. The average segmented PET volume was within 8.0% of the CT volumes. These combined methodologies as outlined above are expected to decrease the conformality index of the tumor dose (open full item for complete abstract)

Master of Science in Engineering (MSEgr), Wright State University, 2012, Electrical Engineering

Given the UHF bands properties of foliage and round penetration, a UHF-SAR image contains both above- and below-surface scatterers. The problem of detecting sub-surface objects is problematic due to the presence of above-surface scatterers in the detection images. In case of a single-pass anomaly image or a two-pass change image, the resulting anomalies or changes are due to scatterers above and below the surface, where the above surface anomalies/changes act as confusers. LIDAR digital elevation models (DEM) provide georegistered information about the above-surface objects present in the UHF-SAR scene. Detection of the above-surface objects in the LIDAR domain is used to rule out above-surface false-alarms in the UHF-SAR domain detection images. A complementary sensor fusion algorithm is implemented which exploits the limited ground penetrating capabilities of UHF-SAR and the false-alarm removal using LIDAR. For unitemporal and multitemporal UHF-SAR collections (both containing multiple-passes and multiple- polarizations) anomaly detection and change detection are implemented, respectively. In this thesis, various pixel-based and feature-based change detection algorithms are implemented to study the effectiveness of multitemporal change detection algorithms. In addition, incorporation of UHF-SAR multiple-passes and multiple-polarizations further improves detection results. The algorithms are tested using data collected under JIEDDOs Halite-1 program, which provides both UHF-SAR and LIDAR DEM.

MS, University of Cincinnati, 2024, Engineering and Applied Science: Computer Science

This thesis explores the potential of multimodal neuroimaging techniques, particularly Electroencephalography (EEG) and Functional Magnetic Resonance Imaging (fMRI), for the classification of visual stimuli. EEG provides excellent temporal resolution, capturing the fast-changing dynamics of brain activity, while fMRI excels in spatial resolution, offering detailed insight into the brain's spatial activation patterns. This dual-fold thesis encompasses both a comprehensive survey and an experimental study. The first part presents a survey of multimodal neuroimaging techniques, focusing on fMRI, EEG, and their integration for improved neural activity analysis, evaluating current approaches and highlighting challenges, limitations, and opportunities in the fusion of these modalities for cognitive and clinical neuroscience applications. The second part involves an experimental study where EEG and fMRI data were fused to classify visual stimuli in a face recognition task. EEG's fine-grained temporal resolution was aligned with fMRI's detailed spatial resolution through temporal matching and feature concatenation. By combining temporal dynamics from EEG and spatial patterns from fMRI, classification performance was enhanced. The fused data classified visual stimuli into three categories: familiar faces, unfamiliar faces, and scrambled faces. Various neural network architectures were tested to capture both temporal and spatial information. The results demonstrate that integrating EEG and fMRI improves the accuracy of visual stimuli classification over single-modality approaches, showcasing the potential of multimodal neuroimaging for a more robust and comprehensive analysis of brain activity. Finally, this study highlights the importance of testing the proposed fusion approach on larger datasets to further validate its effectiveness and explore the generalizability of the results across broader contexts.

Master of Science, The Ohio State University, 1962, Graduate School

Master of Science (MS), Bowling Green State University, 2024, Applied Geospatial Science

Cities around the globe are undergoing significant transformations due to rapid urbanization, fueled by factors such as population growth, economic development, and the migration of people from rural to urban areas. By 2050, an estimated two-thirds of the global population will reside in urban areas, posing significant challenges for sustainable development. Remote sensing data, combined with machine learning modeling approaches play a crucial role in monitoring and analyzing urban sprawl. This study investigates the potential of a machine learning (ML) classification algorithm coupled with data fusion remote sensing techniques to improve land-use and land-cover (LULC) change detection in Kiambu County, Kenya, for the period 2000-2022. It utilizes Landsat data from 2000 to 2022, augmented by Harmonized Landsat Sentinel-2 (HLS) and Sentinel-1 SAR data from 2013 to 2022, for urban land use/land cover (LULC) change detection. Google Earth Engine (GEE) facilitated preprocessing and analysis, refining Synthetic Aperture Radar (SAR) imagery and employing Random Forest (RF) for classification. Integrating Landsat 8/HLS and SAR data enhanced classification accuracy, supported by feature selection, hyperparameter tuning, and spectral band ratios to mitigate data errors. Key indices like NDBI, NBR2, BSI, NDWI, NDVI, and SAVI were crucial for classifying land cover types. From 2000 to 2022, Landsat-based analysis shows significant urbanization. Urban areas grew from 17.8% in 2000 to 22.4% by 2005, 25.7% in 2010, 29.6% in 2015, and 31.9% by 2022. Specifically, for 2015, using Landsat 8 alone, urban areas covered 23.4% (594.0 km²), while fusing Landsat 8 with SAR data raised this to 28.7% (729.4 km²) with improved testing accuracy of 91.7% and validation accuracy of 87.5%. Integrating optical data (HLS and Landsat 8) with SAR and applying ML techniques on GEE, the classification accuracy improved by 5.7% compared to optical data alone. Overall, urbanization in Kiambu Co (open full item for complete abstract)

Master of Science, The Ohio State University, 2024, Electrical and Computer Engineering

As we transition from 2D to 3D information in computer technology, the need for accurate 3D models becomes increasingly important. However, creating these models, especially from point clouds, is challenging due to noise in data and different types of errors in estimating the position and orientation of cameras, which leads to systematic errors among triangulated point clouds among different stereo models. This thesis introduces a method to improve the accuracy of fused point cloud. Our approach, centered on pose graph optimization, utilizes a mix of 3D point clouds from dense reconstructions and 2D matches from original images to calculate the relative pose between point clouds. In this procedure, multiple relative poses are calculated between two point clouds through different paths, bias averaging was introduced to determine the ultimate relative between the two views. Then pose graph is built and point clouds are adjusted to correct systematic biases. Further extending to pose optimization, we explored pose graph optimization for adjusting camera positions. This exploration revealed limitations in initial pose estimates that could lead to inaccuracies in the dense reconstruction. These findings point to a clear direction for future research, focusing on improving pose adjustment techniques to address challenges such as stereo matching noise, outliers, and varied geometric errors.

Master of Science in Electrical Engineering, University of Dayton, 2024, Electrical Engineering

This thesis presents a novel method to control an industrial robotic arm using multiple sensors. This system consists of a hybrid brain activity and vision sensors that convey a human being's intention and visual perception. We fuse and analyze the data from those sensors using a machine learning-based approach to automatically guide the manipulator to a designated location. We believe that this Brain–Machine–Interface (BMI) can greatly alleviate the burdensome traditional method used to program a robot (greatly aids the end-user). We experiment with different modular configurations for the brain activity information, i.e., parallelized models and what we refer to as a global model for fusing the information. We explore various machine learning and pattern recognition techniques as well as existing feature selection methods. Our experimental results show that the subject can control the robot to a destination of interest using a machine—robot–interface. We attain accuracy in the order of 99.6% when it comes to the desired motion and 99.8% for the case of deducing the desired characteristic (color) of the targeted object. These results outperform any similar existing approaches that we have researched. Moreover, in comparison to those similar operational systems, we present a unique modular configuration for brain activity interpretation and object detection mechanism that yields an overall system that is highly computationally efficient. Although, in this work, we implemented and demoed our approach using a simple pick and place demo, our work presents the basic structure underlying a system that can be efficiently used to benefit people with restricted ability to function physically (tetraplegic patients), and allowing them to perform complex and robotics related duties in an industrial setting.

PHD, Kent State University, 2024, College of Arts and Sciences / Department of Computer Science

With the increasing adoption of multi-camera setups for comprehensive monitoring, such as trauma rooms in hospitals, privacy leakage in video surveillance systems has become a significant concern.This research aims to develop an intelligent video surveillance system that leverages skeletal based methods to ensure privacy protection while enabling accurate action recognition. I address the challenge of privacy protection by detecting and blurring human heads or replacing individuals in the system with skeleton representations, while simultaneously applying weighted fusion methods to enhance the action recognition.The main challenge in existing skeletal-based action recognition algorithms is about the insufficient accuracy. To address this issue, I propose two novel algorithms to fuse the results from multiple cameras. The first idea is to divide the floor into many grids, then give them with different fusion weights according to the performance of skeletons in different grids. The second method involves integrating with the YOLO system to intelligently recognize the orientation of individuals within the camera's field of view, then different weights are assigned to the same person when they are captured from different camera perspectives.Those fusion approaches seeks to obtain more reliable and precise action recognition results. By fusing data from various angles, the system enhances the robustness of action recognition, making it suitable for real-world applications. The research also involves extensive experimentation and data analysis to evaluate the proposed algorithm's performance and compare it with existing methods. I aim to achieve a significant improvement in the accuracy of action recognition while ensuring the protection of individuals' privacy in the surveillance context.

MS, University of Cincinnati, 2024, Engineering and Applied Science: Computer Science

Camera traps have become integral tools in wildlife conservation, providing non-intrusive means to monitor and study wildlife in their natural habitats. The utilization of object detection algorithms to automate species identification from Camera Trap images is of huge importance for research and conservation purposes. However, the generalization issue where the trained model is unable to apply its learnings to a never-before-seen dataset is prevalent. This thesis explores the enhancements made to the YOLOv8 object detection algorithm to address the problem of generalization. The study delves into the limitations of the baseline YOLOv8 model, emphasizing its struggles with generalization in real-world environments. To overcome these limitations, enhancements are proposed, including the incorporation of a Global Attention Mechanism (GAM) module, modified multi-scale feature fusion, and Wise Intersection over Union (WIoUv3) as a bounding box regression loss function. A thorough evaluation and ablation experiments reveal the improved model's ability to suppress the background noise, focus on object properties, and exhibit robust generalization in novel environments. The proposed enhancements not only address the challenges inherent in camera trap datasets but also pave the way for broader applicability in real-world conservation scenarios, ultimately aiding in the effective management of wildlife populations and habitats.

Doctor of Philosophy, University of Akron, 2024, Mechanical Engineering

Processing-related defects such as porosity, residual stress, and surface roughness are the primary impediments to the widespread adoption of additive manufacturing in high-performance aerospace structures, primarily in applications where fatigue is an area of concern. Strengthening the surface through an emerging surface treatment approach has the potential to mitigate these defects and subsequently improve the surface quality, as well as increase the fatigue strength of the additively manufactured components. The core objective of this research work was to employ a severe surface plastic deformation (SSPD) process to improve the surface and fatigue properties of additively manufactured Ti-6Al-4V alloys with a particular emphasis on directed energy deposition (DED) re-paired and fully produced electron beam powder bed fusion (EB-PBF), via combination of laser heating (LA) and ultrasonic nanocrystal surface modification (UNSM). Laser heating plus ultrasonic nanocrystal surface modification is an innovative mechanical sur-face treatment tool, and it has been demonstrated as an interesting laser-based mechanical surface treatment technology to induce thicker deformation layer on the surface using low energy input, impact load, low amplitude, and high ultrasonic frequency, leading to enhancement of the microstructure features, surface strength, and resultant mechanical properties of metallic materials. Physical and mechanical characteristics changes in target materials were investigated using optical (OM) and scanning electron microscopy (SEM), X-ray diffraction (XRD), profilometry, and a hardness tester. The results revealed that the proper thermal and impact energies of the applied surface treatment was effective in inducing higher plasticity flow and promoted greater surface grain refinement. Strengthening of metallic alloys through grain refinement is evidenced by achieving maximum strength, a phenomenon referred to as the Hall-Perch principle. In particular, the s (open full item for complete abstract)