Industry Projects
Led the design and validation of an AI segmentation module for an FDA-cleared robotic system. Delivered full treatment planning segmentation pipeline including centerline extraction, 3D mesh generation, and anatomical landmark detection across complex CT anatomies.
Spearheaded development of a cardiac MRI solution from classic methods to deep learning. Achieved high Dice scores for LV, RV, and myocardium segmentation, enabling automatic EF, SV, and CO computation across hospital deployments.
AI Research Projects
HippoQuant: AI powered hippocampal volume quantification system
Develop an AI-powered segmentation model that segments the right hippocampus from cropped MRI volumes and computes its volume. This measurement contributes vital information for monitoring Alzheimer's disease progression, supporting clinicians in decision-making and therapy management.
PneumoAssist: AI-Powered Clinical Decision Support
The goal of this project is to build a deep learning model that can classify chest X-ray images to detect the presence or absence of pneumonia. The intended outcome is to develop a model with diagnostic performance comparable to that of human radiologists, and to prepare the pipeline for submission to the FDA as a software medical device under the 510(k) clearance pathway.
Develop a deep learning regression model to predict the expected hospitalization time for patients using electronic health records (EHR) data. The model helps identify ideal candidates for clinical trials of a new diabetes drug that requires extended inpatient monitoring.
