EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

WICA-Net-M: MRI-Based Brain Tumour Classification Using a Lightweight Wavelet-Integrated Coordinate Attention Network with Frequency-Aware Learning

Md Ashik Khan, Abu Saleh Musa Miah, Md Abdur Rahim, Jungpil Shin, Mohd Nizam Husen

Computers · 04.09.2026 · 15 (9) · 586

DOI: 10.3390/computers15090586 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computers, online erschienen am 04.09.2026. Verfasst von Md Ashik Khan und 4 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background/Objectives: Reported performance on public brain tumour MRI benchmarks is hard to interpret because of near-duplicate train/test overlap, ImageNet pretraining bias, and single-seed evaluation. We address this with a leakage-aware evaluation protocol and a compact model trained entirely from scratch.

Methods

WICA-Net-M is a 2.47 M-parameter CNN whose gated Haar Discrete Wavelet Transform (DWT) separates low- and high-frequency components and fuses them through a learnable gate, complemented by Coordinate Attention. We evaluate it on the standard and image-level deduplicated splits of the Nickparvar brain tumour MRI dataset under a three-seed, leakage-aware protocol, benchmark it against five ImageNet-pretrained baselines and conduct a near-duplicate overlap audit against BRISC 2025.

Results

WICA-Net-M reaches 99.42 ± 0.16% accuracy on the standard V1 split and 95.25 ± 0.32% accuracy/95.17 ± 0.31% macro F1 on the deduplicated V2 split, closely matching five ImageNet-pretrained baselines (95.17–95.67%) with fewer parameters, with sub-half-point differences across three seeds indicating comparable rather than superior accuracy. The audit identifies 861 exact SHA-256 pairs involving 857 of the 1000 BRISC test images against the full Nickparvar collection. Alongside perceptual-hash candidate pairs, this exact overlap shows that BRISC cannot serve as independent external validation.

Conclusions

The descriptive 4.17-point V1-to-V2 difference reflects the combined effects of duplicate removal, class rebalancing, and altered sample composition, with none separable from public releases, though it shows that a scratch-trained compact model can approach pretrained performance on the controlled split. Patient-level leakage remains unresolved because patient identifiers are unavailable. Leakage-aware, multi-seed evaluation should be standard before clinical translation.

Zitieren

APA

Khan, M. A., Miah, A. S. M., Rahim, M. A., Shin, J., & Husen, M. N. (2026). WICA-Net-M: MRI-Based Brain Tumour Classification Using a Lightweight Wavelet-Integrated Coordinate Attention Network with Frequency-Aware Learning. Computers, 15(9), 586. https://doi.org/10.3390/computers15090586

Vancouver

Khan MA, Miah ASM, Rahim MA, Shin J, Husen MN. WICA-Net-M: MRI-Based Brain Tumour Classification Using a Lightweight Wavelet-Integrated Coordinate Attention Network with Frequency-Aware Learning. Computers. 2026;15(9):586. doi:10.3390/computers15090586

BibTeX
@article{Khan2026,
  author = {Khan, Md Ashik and Miah, Abu Saleh Musa and Rahim, Md Abdur and Shin, Jungpil and Husen, Mohd Nizam},
  title = {WICA-Net-M: MRI-Based Brain Tumour Classification Using a Lightweight Wavelet-Integrated Coordinate Attention Network with Frequency-Aware Learning},
  journal = {Computers},
  year = {2026},
  volume = {15},
  number = {9},
  pages = {586},
  doi = {10.3390/computers15090586}
}
Kostenlos starten →