Enhancing Deep Learning-Based Liver Vessel Segmentation on MRI with Image Translation Techniques
Motivation: To improve liver vessel segmentation on MRI under annotation constraints.Goal(s): Apply an advanced unpaired image translation technique, SynDiff, to create synthetic MR images from CT data.Approach: By incorporating vessel masks in the translation process, the optimized SynDiff models generated synthetic images that facilitated more effective pretraining of segmentation models.Results: Validated across multiple pretraining settings, the refined SynDiff approach surpassed the standard nnU-Net and other pretraining-based methods, substantially improving liver vessel segmentation performance.Impact: This study remarkably advances liver vessel segmentation on MRI, demonstrating that synthetic data can effectively augment limited datasets, leading to improved model performance. It has great potential for broader applications in medical image analysis.
Authors
- Brian Toner (ORCID: https://orcid.org/0000-0002-3662-4689)
- Stephan Kannengießer (ORCID: https://orcid.org/0000-0001-6345-9132)
- Sabine Mollus
- Deniz Karakay
- Ali Bilgin (ORCID: https://orcid.org/0000-0003-4196-4036)
- Sevgi Gökçe Kafalı (ORCID: https://orcid.org/0000-0001-5941-5399)
- Timoteo Delgado (ORCID: https://orcid.org/0000-0001-6709-410X)
- Holden H. Wu (ORCID: https://orcid.org/0000-0002-2585-5916)
- María I. Altbach (ORCID: https://orcid.org/0000-0001-9368-2458)
- Eze Ahanonu (ORCID: https://orcid.org/0009-0004-5996-3589)
- Vibhas Deshpande (ORCID: https://orcid.org/0000-0003-0264-6553)
- Saša Grbić (ORCID: https://orcid.org/0000-0002-7564-9811)
- Yanbo Zhang (ORCID: https://orcid.org/0000-0002-7627-488X)
- Wenqi Zhou (ORCID: https://orcid.org/0009-0000-3197-595X)
Publication Details
- Journal
- Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition
- Published
- 2024-11-26
- DOI
- https://doi.org/10.58530/2024/1398
- Primary Topic
- Brain Tumor Detection and Classification
- Type
- article
- Field-Weighted Citation Impact
- 0.00