SERA-Net: Rethinking CNN Design for Brain Tumor Classification via Squeeze-and-Excite Attention and Residual Learning
Jyoti Ranjan Sahoo, Surendra Kumar Nanda, Ganapati Panda, Ashutosh Soni, Jayanti Rout, Manob Jyoti Saikia · 2026
Read SERA-Net: Rethinking CNN Design for Brain Tumor Classification via Squeeze-and-Excite Attention and Residual Learning by Jyoti Ranjan Sahoo, Surendra Kumar Nanda, Ganapati Panda, Ashutosh Soni, Jayanti Rout, Manob Jyoti Saikia (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/pmc-13603351.txt.
Description
Jyoti Sahoo and five colleagues present a 2026 paper on a brain tumor scanner built from scratch rather than borrowed. Most classification networks start from photo-trained backbones and retrain only a small head, leaving most weights unable to adapt to MRI scans. Their network, SERA-Net, trains fully on the medical images themselves.
Four blocks mix channel attention, which reweights informative image features, with residual shortcuts that keep training stable. An ablation study added parts one at a time to prove each earns its place. With only about five million parameters, the network scored about ninety-five point four percent accuracy and beat six pretrained rivals by five to ten points. The authors offer the design recipe as much as the model.
Catalog record
| CREATOR | Jyoti Ranjan Sahoo, Surendra Kumar Nanda, Ganapati Panda, Ashutosh Soni, Jayanti Rout, Manob Jyoti Saikia |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |