PhD Candidate
AI Specialist
Data scientist
Software Engineer
PhD Candidate
AI Specialist
Data scientist
Software Engineer
1. Bifurcation Modeling
Since April 2023
I am developing novel methods for detecting and analyzing bifurcations in both airways and arteries within thoracic CT scans, aiming to improve disease understanding and treatment planning. This project includes creating and releasing BifDet, the first public dataset for 3D airway bifurcation detection.
- Keshavarzi, A., Bouniot, Q., Smith, B. M., & Angelini, E. (2024). “BifDet: A 3D Bifurcation Detection Dataset for Airway-Tree Modeling”, Under Review for NeurIPS 2024.
2. Tubular Structures Segmentation
Since Nov. 2022
I am focusing on incorporating prior knowledge and loss function regularization techniques into deep learning models to achieve topologically correct predictions for tubular structures like airways and arteries in lung CT scans. This research aims to improve the accuracy and reliability of segmentation results for these critical anatomical structures.
- Keshavarzi, A., & Angelini, E. (2024). “Few-Shot Airway-Tree Modeling using Data-Driven Sparse Priors”, 21st IEEE International Symposium on Biomedical Imaging (ISBI). Athens, Greece. LINK
- Jacovella, M., Keshavarzi, A., & Angelini, E. (2024). “Curriculum Learning for Few-Shot Domain Adaptation in CT-based Airway Tree Segmentation”. LINK
3. Anatomical-Tree Modeling
Since May. 2024
Building upon the BifDet dataset, I am developing comprehensive airway and arterial tree models to advance our understanding of respiratory and cardiovascular health. This research aims to provide detailed insights into the morphology and bifurcation parameters of both systems, potentially informing personalized treatment strategies.