MONICA
A web application for automated whole–optic-nerve contour extraction and morphometric analysis, validated across taxonomic orders and image-quality levels. MONICA takes a histologic optic-nerve image and returns reproducible, quantitative morphometrics without manual tracing.
AxoClassNet
A deep-learning model that segments and classifies healthy versus degenerating (necrotic) optic-nerve axons from histology, enabling automated, quantitative assessment of optic-nerve health at scale.
Automated optic-nerve axon quantification
An adapted AxonDeepSeg pipeline that quantifies axon counts, axon density, and parenchymal area in mouse optic-nerve cross-sections — together with an independent benchmark comparing multiple machine-learning tools, which found limited generalizability upon external validation.
AI for glaucoma detection (fundus & OCT)
Deep-learning systems for glaucoma detection from color fundus photographs and OCT, including image-quality and gradeability models, generative image enhancement, and evaluations of multimodal foundation models. This work spans first-author publications in Ophthalmology Science, Translational Vision Science & Technology, and JAMA Ophthalmology.
Tools & methods
Languages: Python, R, C/C++, Java, MATLAB
ML / CV: deep learning for medical imaging — segmentation, classification, foundation-model fine-tuning & evaluation, multimodal models
Focus: reproducible pipelines, independent validation, and translating research methods into deployable web tools