Exploiting Label-Independent Regularization from Spatial Patterns for Whole Slide Image Analysis
Weiyi Wu, Xinwen Xu, Chongyang Gao, Xingjian Diao, Siting Li, Jiang Gui · 2026
Read Exploiting Label-Independent Regularization from Spatial Patterns for Whole Slide Image Analysis by Weiyi Wu, Xinwen Xu, Chongyang Gao, Xingjian Diao, Siting Li, Jiang Gui (2026) 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-2602-19487.txt.
Description
Whole slide images offer gigapixel views of tissue, but their size and scarce labels hinder analysis. Weiyi Wu and five co-authors note that multiple instance learning faces a basic mismatch: one bag-level label must guide learning across many patch-level features. Sparse labels make it hard to spot discriminative patches reliably, leading to unstable tuning and poor solutions.
The team offers a spatially tuned framework that uses spatial links among patches as tuning signals that need no labels. A shared representation space is learned by jointly tuning spatial reconstruction from features and label-guided classification, keeping consistency between structural patterns and label signals. Tests on multiple public data sets show clear gains for whole slide image analysis.
Catalog record
| CREATOR | Weiyi Wu, Xinwen Xu, Chongyang Gao, Xingjian Diao, Siting Li, Jiang Gui |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-19487