Scholarship @ UWindsor

Scholarship @ UWindsor is the institutional repository of the University of Windsor (UWindsor), showcasing and preserving the UWindsor community’s scholarly outputs, as well as items from the Leddy Library’s Archives & Special Collections. Its mission is to disseminate and preserve knowledge created or housed at the University of Windsor.

Contact scholarship@uwindsor.ca for more information.

Communities in Scholarship @ UWindsor

Select a community to browse its collections.

Now showing 1 - 4 of 4

Conferences and Conference Proceedings 3167
Papers, presentations and abstracts of conferences held at the University of Windsor, in person and virtually.
Digitized Local Collections 1552
Digitized local items from the collections of the Leddy Library, University of Windsor, and community partners.
Faculties, Departments and Research Units 3205
Open Access Faculty publications, reports and working papers from academic departments at the University of Windsor.
Theses, Dissertations, and Major Papers 9579
Formal graduate original research from the University of Windsor's Masters and Doctoral programs.

Recent Submissions

Access status: Open Access ,
Cadmium spurred Microcystis aeruginosa to unleash more toxic metabolites
(Academic Press, 2025-02-21) Zhou, Yuan a, b ; a ; a ; c ;; Li, Jingjing; Pei, Ying; Xu, Runbing; Zi, Jinmei; Harshaw, Keira; Chang, Xuexiu
Cyanobacterial harmful algal blooms (cHABs), normally dominated by Microcystis aeruginosa, pose a threat to aquatic ecosystems due to the release of various harmful metabolites. Cadmium (Cd), a heavy metal commonly found in surface water and sediments, often coincides with cHABs in eutrophic lakes. However, the ecotoxicological effects of Cd on M. aeruginosa and the potential for combined toxicity are not yet fully understood. In this study, we determined the effective concentrations of cadmium from 10 % (EC10) to 50 % (EC50) for M. aeruginosa based on cell density inhibition. We then conducted a combined analysis focusing on the impact of a low dose Cd (EC10, 139 μg/L) on the physiological factors, transcriptome and both intracellular and extracellular metabolites of M. aeruginosa. We found that Cd treatment decreased M. aeruginosa chlorophyll a content by 24.5 %, which coincided with the suppression of genes linked to ribosomal and photosynthesis pathways. However, Cd exposure stimulated the synthesis and extracellular release of cellular compounds by enhancing amino acid and carbohydrate metabolism. This led to elevated extracellular levels of amino acids, organic acids, and secondary metabolites - including peptides, lipids, benzenoids, terpenes, sterols, and glycosides - which could serve as potential toxic metabolites of cyanobacteria. These changes were driven by the activation of osmoregulatory mechanisms, antioxidant-related amino acids, and ATP-binding cassette transport and secretion systems. Our research indicated that low Cd concentrations could stimulate the synthesis and release of toxic metabolites and exacerbate cHAB threats in eutrophic lakes, underscoring the importance of addressing multiple stressors in freshwater environments.
Access status: Open Access ,
Visual-language transformer-based tomato leaf disease detection for portable greenhouse monitoring device
(BioMed Central Ltd, 2025-10-28) Kaur, Manveen; Singh, Rajmeet; Alirezaee, Shahpour; Hussain, Irfan
Tomato leaf diseases pose a significant threat to global food security, necessitating accurate and efficient detection methods. This paper introduces the Tomato Leaf Disease Visual Language Model (TLDVLM), a novel approach based on the BLIP-2 architecture enhanced with Low-Rank Adaptation (LoRA), for precise classification of 10 distinct tomato leaf diseases. Our methodology integrates a sophisticated image preprocessing pipeline, utilizing GroundingDINO for robust leaf detection and SAM-2 for pixel-level segmentation, ensuring that the model focuses solely on relevant plant tissue. The TLDVLM leverages the powerful multimodal understanding of BLIP-2, with LoRA applied to its Q-Former module, enabling parameter-efficient fine-tuning without compromising performance. Comparative experiments demonstrate that the TLDVLM significantly outperforms baseline models, including CLIP-LoRA and ConvNeXT-tiny, achieving an accuracy of 97.27%, a precision of 0.9587, a recall of 0.9789, and an F1-score of 0.9681. Beyond classification, the finetuned TLDVLM checkpoints are integrated into a practical application for new image inference. This application displays the raw and segmented images, the predicted disease, and offers functionalities to fetch comprehensive information on disease causes and remedies using external APIs (e.g., OpenAI), with an option to download a PDF summary for offline access on a portable device. This research highlights the potential of LoRA-adapted Vision-Language Models in developing highly accurate, efficient, and user-friendly agricultural diagnostic tools.
Access status: Open Access ,
Assessing Driver Comfort and In-Chair Movements in an Automotive Seat with Anti-Fatigue Massage Function
(MDPI, 2025-05-16) Behzad, Adele a Send mail to Behzad A. ; a Send mail to Michael I. ; a Send mail to Eniyandunmo D. ; b Send mail to Andrews D.M. ;; Michael, Ifeoma; Eniyandunmo, David; Andrews, David M.; Cole, Shane; Arthurs, Katherine L.; Kim, Eunsik
Prolonged driving is associated with fatigue and reduced comfort, jeopardizing driver safety. This study proposes an innovative Anti-Fatigue Massage Function (AFMF) system integrated into a driver’s seat to improve subjective comfort and decrease compensatory movements during extended driving due to fatigue. In total, 24 participants (12 males, 12 females) completed two 60 min simulated driving sessions—one with the AFMF activated and one with it deactivated. Subjective comfort was measured every 10 min using a 5-point Likert scale, while objective In-Chair Movements (ICMs) were manually recorded by expert researchers from dual-camera recordings. ART ANOVA revealed that the AFMF-equipped seat significantly enhanced comfort ratings and reduced ICM frequencies compared to the deactivated condition. These findings suggest that the AFMF system can enhance driver well-being and mitigate fatigue-related risks during prolonged driving.
Access status: Open Access ,
Social dimensions of shark–human interactions in a large remote marine protected area
(John Wiley & Sons Ltd.) Collins, Claire; Weber, Sam; Clarke, Lucy; Gollock, Matthew; Hussey, Nigel E.; Simpson, Daniel; Simpson, Tiffany; Curnick, David
The impact of shifting marine conservation policy, including marine protected area (MPA) designation, in shaping interactions between humans and imperilled species, such as sharks, remains understudied, despite its importance in determining the success of these interventions. We investigated perceptions of shark–human interactions (SHI) among the community in a remote, large-scale MPA (Ascension Island) where two recent shark attacks and perceived general increases in interactions (mostly with Galapagos and silky sharks), including depredation in recreational fisheries, have occurred. From 2023 to 2024, informal semistructured interviews (n = 34) were conducted with island residents and analysed using two theory-driven thematic frameworks to understand the level and drivers of conflict. We showed considerable social impact of SHI, including reduced human well-being and substantial lifestyle changes, with both depredation and human attacks driving dispute-level conflict. Strong social and familial connections on island resulted in narratives around attacks persisting and trauma resulting from attacks drove heightened perceived risk. Underlying conflict was further exacerbated by the perceived recurrent and unpredictable nature of negative SHI compounded with the interactions being perceived as abnormal with limited information on socio-ecological drivers. Some felt excessive chumming by historic recreational fisheries, mostly engaged in by non-residents, had also involuntarily exposed them to heightened risk by increasing SHI. There was also no consensus of what shark species were behind the increased interactions. Management resolutions were perceived as minimal but were not widely viewed as negative. However, divergent views on the use of lethal control and the need for conservation measures, such as banning shark exploitation, were evident. A key theme emerged around the need for wider community participation in the research and management processes. Policy implications. Our results highlight the critical importance of demystifying marine species, particularly in terms of understanding socio-ecological drivers of human–wildlife interactions, to combat escalation into human–wildlife conflict. This is particularly important to maintain support for large-scale MPAs and species-specific conservation. Read the free Plain Language Summary for this article on the Journal blog.
Access status: Open Access ,
Hierarchical Multi-Scale Patch Attention and Global Feature-Adaptive Fusion for Robust Occluded Face Recognition
(Institute of Electrical and Electronics Engineers Inc., 2025-06-24) Hejazi, Elhamsadat; Ahmadi, Majid; Ahmadi, Arash
Occluded face recognition remains a challenging problem in biometric identification, where real-world obstructions such as masks, sunglasses, scarves, and hands obscure key facial features. To address this, we introduce a dual-branch architecture that combines a Local Multi-Patch Attention Module (LMPAM) for extracting localized features with a Global Self-Attention Channel Module (GSACM) to enhance overall feature representation. The local branch utilizes Multi-Scale Patch Attention to adaptively emphasize visible facial regions, ensuring robust feature learning from unoccluded areas. Meanwhile, the global branch employs Self-Attention with Channel Recalibration to enhance discriminative features, capturing long-range dependencies while suppressing occlusion-induced noise. The two branches are integrated using Dynamic Weighted Local-Global Fusion (DW-LG), allowing the model to balance local and global information effectively. Unlike predefined occlusion-aware methods, our approach generalizes across occlusions of varying types, regions, and sizes and demonstrates robustness on multiple datasets with changes in illumination, pose, and facial expression—without requiring explicit localization. Extensive evaluations on CASIA-WebFace, LFW, and AR datasets demonstrate the effectiveness of our approach, achieving higher recognition performance under severe occlusion conditions.