Renaud Duval
Research Interests
Make your next move.
Prepare your application
Review drafts before you send them.
Explore related searches
Articles (7)
Year:
2026
Year:
2026
Year:
2026
DNA fiber assays are powerful tools for investigating replication dynamics at the single-molecule level. However, their application and widespread adoption has been hampered by the labor-intensive and tedious nature of manual analysis of large numbers of images. Quantification of labeled DNA fibers typically depends on subjective examination, selection, and annotation of individual fibers from fluorescence microscopy images reducing inter-user consistency, reproducibility, and experimental throughput. To address these issues, we developed DNAi, a computer vision tool based on deep learning allowing automated detection and quantification of labeled DNA fiber length. DNAi was trained on a large and diverse dataset of manually annotated images of DNA fibers and matches human performance and accuracy in segmentation and length measurement across a wide range of experimental conditions. The open-source tool includes a user-friendly interface, which permits visual validation and manual selection of segmented fibers. Overall, DNAi enables robust, rapid, and reproducible DNA fiber analysis, and is freely available. Graphical abstract
Year:
2025
Purpose To compare the performance and cost-effectiveness of DeepSeek-R1 with OpenAI o1 in diagnosing and managing ophthalmology clinical cases. Study Design Cross-sectional evaluation. Methods A total of 300 clinical cases spanning 10 different ophthalmology subspecialties were collected from StatPearls. Each case presented a multiple-choice question regarding the diagnosis or management of the clinical case. DeepSeek-R1 was accessed through its public chat-based interface, while OpenAI o1 was queried via an Application Program Interface (API) with a standardized temperature setting of 0.3. Both models were prompted using the Plan-and-Solve+ (PS+) prompt engineering method, instructing them to answer multiple choice questions for each case. Performance was calculated as the proportion of correctly answered multiple choice questions. McNemar’s test was employed to compare the two models’ performance on paired data. Inter-model agreement for correct diagnoses was evaluated via Cohen’s kappa. A token-based cost analysis was performed to estimate the comparative expenditures of running each model at scale, accounting for both input prompts and model-generated output. Results DeepSeek-R1 and OpenAI o1 both achieved an identical overall performance of 82.0% (n=246/300; 95% CI: 77.3-85.9). Subspeciality-specific analysis revealed numerical variation in performance, though none of these comparisons reached statistical significance (p>0.05). Agreement in performance between the models was moderate overall (κ=0.503, p<0.001), with substantial agreement in Refractive Management/Intervention (κ=0.698, p<0.001) and moderate agreement in Retina/Vitreous (κ=0.561, p<0.001) and Ocular Pathology/Oncology (κ=0.495, p<0.01) cases. Cost analysis indicated an approximately 15-fold reduction in per-query, token-related expenses when using DeepSeek-R1 compared with OpenAI o1 for the same workload. Conclusions DeepSeek-R1 demonstrates robust diagnostic reasoning and management decision-making capabilities, performing comparably to OpenAI o1 across a range of ophthalmic subspecialty cases, while also offering a substantial reduction in usage costs. These findings highlight the feasibility of utilizing open-weight, reinforcement learning-augmented LLMs as an accessible, cost-effective alternative to proprietary models.
Year:
2025
Year:
2024
Background Code-free deep learning (CFDL) is a novel tool in artificial intelligence (AI). This study directly compared the discriminative performance of CFDL models designed by ophthalmologists without coding experience against bespoke models designed by AI experts in detecting retinal pathologies from optical coherence tomography (OCT) videos and fovea-centered images. Methods Using the same internal dataset of 1,173 OCT macular videos and fovea-centered images, model development was performed simultaneously but independently by an ophthalmology resident (CFDL models) and a postdoctoral researcher with expertise in AI (bespoke models). We designed a multi-class model to categorize video and fovea-centered images into five labels: normal retina, macular hole, epiretinal membrane, wet age-related macular degeneration and diabetic macular edema. We qualitatively compared point estimates of the performance metrics of the CFDL and bespoke models. Results For videos, the CFDL model demonstrated excellent discriminative performance, even outperforming the bespoke models for some metrics: area under the precision-recall curve was 0.984 (vs. 0.901), precision and sensitivity were both 94.1% (vs. 94.2%) and accuracy was 94.1% (vs. 96.7%). The fovea-centered CFDL model overall performed better than video-based model and was as accurate as the best bespoke model. Conclusion This comparative study demonstrated that code-free models created by clinicians without coding expertise perform as accurately as expert-designed bespoke models at classifying various retinal pathologies from OCT videos and images. CFDL represents a step forward towards the democratization of AI in medicine, although its numerous limitations must be carefully addressed to ensure its effective application in healthcare.
Year:
2024
Collaborators (5)
Santiago Costantino
Chercheur
Université de Montréal
Pearse Keane
University College London
Hugo Wurtele
Full Professor (Professeur titulaire)
Université de Montréal
Rajeev H. Muni
-
Andrew Farah
McGill University
Make your next move.
Prepare your application
Review drafts before you send them.