Hypergraph Tversky-Aware Domain Incremental Learning for Brain Tumor Segmentation with Missing Modalities
Junze Wang, Lei Fan, Weipeng Jing, Donglin Di, Yang Song, Sidong Liu, Cong Cong✉ · 2025
Read Hypergraph Tversky-Aware Domain Incremental Learning for Brain Tumor Segmentation with Missing Modalities by Junze Wang, Lei Fan, Weipeng Jing, Donglin Di, Yang Song, Sidong Liu, Cong Cong✉ (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-16809.txt.
Description
Junze Wang and co-authors segment brain tumors when some MRI scan types are missing, in a 2025 paper. Tumor outlining usually wants four MRI sequences, but clinics acquire them in stages and some are always absent for some patients. Retraining a model each time a new sequence arrives wastes effort and can erase old skill, a failure called forgetting.
Their method, ReHyDIL, learns new scan domains step by step while replaying stored examples to protect prior knowledge. A hypergraph links image regions beyond simple pairs, and a Tversky-style loss weights missed tumor against false alarm in a tunable way. Steady accuracy across arriving modalities would make tumor tools fit real hospital workflows.
Catalog record
| CREATOR | Junze Wang, Lei Fan, Weipeng Jing, Donglin Di, Yang Song, Sidong Liu, Cong Cong✉ |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 24 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-16809