PSO-UNet: Particle Swarm-Optimized U-Net Framework for Precise Multimodal Brain Tumor Segmentation Conference: Genetic and Evolutionary Computation Conference; Accepted; as Poster CCS: Computing methodologies Image processing CCS: Computing methodologies Image segmentation CCS: Networks Network arch
Shoffan Saifullah, Rafał Dreżewski · 2025
Read PSO-UNet: Particle Swarm-Optimized U-Net Framework for Precise Multimodal Brain Tumor Segmentation Conference: Genetic and Evolutionary Computation Conference; Accepted; as Poster CCS: Computing methodologies Image processing CCS: Computing methodologies Image segmentation CCS: Networks Network arch by Shoffan Saifullah, Rafał Dreżewski (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-2503-19152.txt.
Description
Shoffan Saifullah and Rafał Dreżewski tune a brain tumor segmentation network with a swarm search instead of hand picking. Their 2025 paper pairs particle swarm optimization with the U-Net architecture for MRI scans. Manual tuning struggles in the wide space of filters, kernel sizes, and learning rates.
The swarm tunes those settings dynamically for each run. Overlap scores reach about 0.96 on one major dataset and 0.95 on another. The model stays lean at under eight million parameters and runs in about fifteen minutes. That speed beats comparable U-Net frameworks by a wide margin, the authors say. Results hold across MRI types and tumor classes. Future work will test hybrid search methods and other nature-inspired optimizers.
Catalog record
| CREATOR | Shoffan Saifullah, Rafał Dreżewski |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-19152