EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Firefly Optimized UNET: A Novel Approach for High-Precision Segmentation of Meningiomas and Gliomas in MRI Scans of the Brain

Narendra Kumar, Filippo Verre, Anurag Aeron

Journal of Artificial Intelligence in Medical Applications · 18.04.2026 · 2 (1) · 16-24

DOI: 10.65511/jaima.v2i1.1115 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Artificial Intelligence in Medical Applications, online erschienen am 18.04.2026. Verfasst von Narendra Kumar und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Cancer is today one of the most significant diseases in the world because of environmental change, contamination, and hereditary. By receiving the right care and early diagnosis, cerebral cancers can be avoided. Magnetic Resonance Imaging (MRI) manual diagnosis is quite challenging for a large patient database since it takes longer. Without the assistance of clinical subject matter experts, advances in PC vision and AI computations can efficiently identify and order cancers from MRI data. The procedure of classifying tumors has been approached in a variety of ways in the past. Direct and indirect pixel-based techniques were created for the segmentation process. A segmentation map is produced directly from the input image using a UNET-based method. In this article, proposed a brand-new UNET approach for categorizing brain tumors using MRI images. The tumor regions are divided into three subregions for classification purposes, including edema, necrosis, and increasing tumor. A redesigned UNET architecture is used for the testing procedure after a highly accurate training method employing mask pictures has been completed. The weight optimization in UNET makes use of FireFly optimization to generate exceptionally accurate classification results. By delivering randomly generated weight values, optimization is done throughout the training phase. This approach achieves 98.34% accuracy for the core brain region with a processing time reduction of 61.24%. This work improves the performance of accuracy better than other existing works.

Zitieren

APA

Kumar, N., Verre, F., & Aeron, A. (2026). Firefly Optimized UNET: A Novel Approach for High-Precision Segmentation of Meningiomas and Gliomas in MRI Scans of the Brain. Journal of Artificial Intelligence in Medical Applications, 2(1), 16-24. https://doi.org/10.65511/jaima.v2i1.1115

Vancouver

Kumar N, Verre F, Aeron A. Firefly Optimized UNET: A Novel Approach for High-Precision Segmentation of Meningiomas and Gliomas in MRI Scans of the Brain. Journal of Artificial Intelligence in Medical Applications. 2026;2(1):16-24. doi:10.65511/jaima.v2i1.1115

BibTeX
@article{Kumar2026,
  author = {Kumar, Narendra and Verre, Filippo and Aeron, Anurag},
  title = {Firefly Optimized UNET: A Novel Approach for High-Precision Segmentation of Meningiomas and Gliomas in MRI Scans of the Brain},
  journal = {Journal of Artificial Intelligence in Medical Applications},
  year = {2026},
  volume = {2},
  number = {1},
  pages = {16--24},
  doi = {10.65511/jaima.v2i1.1115}
}
Kostenlos starten →