Tied Prototype Model for Few-Shot Medical Image Segmentation
Hyeongji Kim (✉), Stine Hansen, Michael Kampffmeyer · 2025
Read Tied Prototype Model for Few-Shot Medical Image Segmentation by Hyeongji Kim (✉), Stine Hansen, Michael Kampffmeyer (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-2506-22101.txt.
Description
Hyeongji Kim, Stine Hansen, and Michael Kampffmeyer presented a 2025 method for marking organs in medical scans when only a few marked scans exist to learn from. Such methods build a sample patch for the target part and treat the rest of the scan as odd or out of place. A past method named ADNet used just one sample patch per class and fixed cut points, which coped poorly as patients and organs change.
The new model, called TPM, ties the sample spots for the target and the rest of the image to one shared base. That tie lets it use more than one sample and more than one class at once. It also sets cut points from how often each class tends to occur, and tests show gains in scan marking scores.
Catalog record
| CREATOR | Hyeongji Kim (✉), Stine Hansen, Michael Kampffmeyer |
|---|---|
| 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/2506-22101