Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound
Precisely estimating lumen boundaries in intravascular ultrasound (IVUS) is needed for sizing interventional stents to treat deep vein thrombosis (DVT). Unfortunately, current segmentation networks like the UNet lack the precision required for clinical adoption in IVUS workflows. This arises due to...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
Massachusetts Institute of Technology
2024
|
| Online Access: | https://hdl.handle.net/1721.1/157219 |
| _version_ | 1824457883100643328 |
|---|---|
| author | Chen, Yiming |
| author2 | Golland, Polina |
| author_facet | Golland, Polina Chen, Yiming |
| author_sort | Chen, Yiming |
| collection | MIT |
| description | Precisely estimating lumen boundaries in intravascular ultrasound (IVUS) is needed for sizing interventional stents to treat deep vein thrombosis (DVT). Unfortunately, current segmentation networks like the UNet lack the precision required for clinical adoption in IVUS workflows. This arises due to the difficulty of automatically learning accurate lumen contour from limited training data while accounting for the radial geometry of IVUS imaging. We propose the Geo-UNet framework to address these issues via a design informed by the geometry of the lumen contour segmentation task, building anatomical constraints directly into the architecture. We first convert the input data and segmentation targets from Cartesian to polar coordinates. Starting from a convUNet feature extractor, we propose a two-task setup, one for conventional pixel-wise labeling and the other for single boundary lumen-contour localization. We directly combine the two predictions by passing the predicted lumen contour through a new activation (named CDFeLU) to filter out spurious pixel-wise predictions. Our unified loss function carefully balances area-based, distance-based, and contour-based penalties to provide near clinical-grade generalization in unseen patient data. We also introduce a lightweight, inference-time technique to enhance segmentation smoothness. The efficacy of our framework on a venous IVUS dataset is shown against state-of-the-art models. We will make the code repository for this project available soon after approval from industry collaborators. |
| first_indexed | 2025-02-19T04:17:04Z |
| format | Thesis |
| id | mit-1721.1/157219 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2025-02-19T04:17:04Z |
| publishDate | 2024 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/1572192024-10-10T04:09:05Z Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound Chen, Yiming Golland, Polina Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Precisely estimating lumen boundaries in intravascular ultrasound (IVUS) is needed for sizing interventional stents to treat deep vein thrombosis (DVT). Unfortunately, current segmentation networks like the UNet lack the precision required for clinical adoption in IVUS workflows. This arises due to the difficulty of automatically learning accurate lumen contour from limited training data while accounting for the radial geometry of IVUS imaging. We propose the Geo-UNet framework to address these issues via a design informed by the geometry of the lumen contour segmentation task, building anatomical constraints directly into the architecture. We first convert the input data and segmentation targets from Cartesian to polar coordinates. Starting from a convUNet feature extractor, we propose a two-task setup, one for conventional pixel-wise labeling and the other for single boundary lumen-contour localization. We directly combine the two predictions by passing the predicted lumen contour through a new activation (named CDFeLU) to filter out spurious pixel-wise predictions. Our unified loss function carefully balances area-based, distance-based, and contour-based penalties to provide near clinical-grade generalization in unseen patient data. We also introduce a lightweight, inference-time technique to enhance segmentation smoothness. The efficacy of our framework on a venous IVUS dataset is shown against state-of-the-art models. We will make the code repository for this project available soon after approval from industry collaborators. M.Eng. 2024-10-09T18:29:06Z 2024-10-09T18:29:06Z 2024-09 2024-10-07T14:34:39.810Z Thesis https://hdl.handle.net/1721.1/157219 Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) Copyright retained by author(s) https://creativecommons.org/licenses/by-nc-nd/4.0/ application/pdf Massachusetts Institute of Technology |
| spellingShingle | Chen, Yiming Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound |
| title | Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound |
| title_full | Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound |
| title_fullStr | Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound |
| title_full_unstemmed | Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound |
| title_short | Geo-UNet: A Geometrically Constrained Neural Framework for Clinical-Grade Lumen Segmentation in Intravascular Ultrasound |
| title_sort | geo unet a geometrically constrained neural framework for clinical grade lumen segmentation in intravascular ultrasound |
| url | https://hdl.handle.net/1721.1/157219 |
| work_keys_str_mv | AT chenyiming geounetageometricallyconstrainedneuralframeworkforclinicalgradelumensegmentationinintravascularultrasound |