Bridging Topology and Deep Representation Learning: A TDA-ViT Fusion Model for Four-Class Brain Tumor Classification
Faisal Ahmed · 2026
Read Bridging Topology and Deep Representation Learning: A TDA-ViT Fusion Model for Four-Class Brain Tumor Classification by Faisal Ahmed (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-2606-00927.txt.
Description
Ahmed fuses shape math with deep vision to sort brain tumors in this 2026 paper. MRI scans guide early diagnosis and care choices for brain tumors. Vision Transformers read such scans well, learning broad context across the image. But they can miss the deep structural patterns inside a tumor's shape.
His fusion model pairs two views of each scan. Topological Data Analysis extracts shape traits: geometry, connectivity, and form. A pretrained Vision Transformer learns high-level semantic traits in parallel. The two feature sets merge into one stronger code for the classifier. Tests use the BRISC2025 dataset, with glioma, meningioma, pituitary tumor, and non-tumor cases. The fusion scores 99.10 percent accuracy, he reports. Precision, recall, and F1 all sit near 99 percent as well. Each view alone does worse than the pair. The work argues that tumor shape is signal, not noise, for scan readers.
Catalog record
| CREATOR | Faisal Ahmed |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.00927