EUVIMEDEUVIMED entdecken
Crossref

ICARECS-2026 - a Hybrid Cnn–transformer Framework for Joint Brain Tumor Segmentation and Multi-Class MRI Classification

Almas Begum, Sajja Bhanu Teja, Chemirthipalli Pradeep, Bhattaru V. S., Harshavardhan Joshi, Koneru Rupesh Reddy, B. Gnanapradeep

i-manager's Journal on Image Processing · 20.06.2026 · 13 (2) · 1

DOI: 10.26634/jip.13.2.11329 ↗

Kurz gefasst

Fachartikel in der Zeitschrift i-manager's Journal on Image Processing, online erschienen am 15.06.2026. Verfasst von Almas Begum und 6 weiteren Autor:innen. Die Arbeit nennt 15 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Diagnosis of brain tumors through magnetic resonance imaging (MRI) plays an important role in the early diagnosis and effective planning of therapy. Nevertheless, there is time-consumption in hand examination of MRI scans, which is subject to inter-clinician variations. This paper will solve this issue by suggesting a Hybrid CNN-Transformer architecture as a joint brain tumor segmentation and multi-class MRI classification model. The proposed model combines the application of CNNs to extract local spatial information with transformer-based attention mechanisms to obtain global contextual relationships of MRI images. The first module is a segmentation module, which is used to identify tumor regions, and the classification network is then able to target important tumor regions. This is followed by the classification of the tumors into four types: glioma, meningioma, pituitary tumor, and non-tumor using the fused CNN and transformer characteristics. Experimental analysis of MRI data sets shows that the suggested model attains the highest performance in the form of 97.8% accuracy, 97.2% precision, 96.8% recall, and 97.0% F1- score, which is better than popular machine learning, CNN-based, and attention-based models. The findings suggest a tremendous enhancement of the accuracy and reliability of tumor classification when segmentation-guided learning is combined with transformer-based contextual modeling. The framework suggested is a good solution to automated brain tumor analysis and can be useful to facilitate clinical decisions in the field of medical imaging.

Zitieren

APA

Begum, A., Teja, S. B., Pradeep, C., S., B. V., Joshi, H., Reddy, K. R., & Gnanapradeep, B. (2026). ICARECS-2026 - a Hybrid Cnn–transformer Framework for Joint Brain Tumor Segmentation and Multi-Class MRI Classification. i-manager's Journal on Image Processing, 13(2), 1. https://doi.org/10.26634/jip.13.2.11329

Vancouver

Begum A, Teja SB, Pradeep C, S. BV, Joshi H, Reddy KR, et al. ICARECS-2026 - a Hybrid Cnn–transformer Framework for Joint Brain Tumor Segmentation and Multi-Class MRI Classification. i-manager's Journal on Image Processing. 2026;13(2):1. doi:10.26634/jip.13.2.11329

BibTeX
@article{Begum2026,
  author = {Begum, Almas and Teja, Sajja Bhanu and Pradeep, Chemirthipalli and S., Bhattaru V. and Joshi, Harshavardhan and Reddy, Koneru Rupesh and Gnanapradeep, B.},
  title = {ICARECS-2026 - a Hybrid Cnn–transformer Framework for Joint Brain Tumor Segmentation and Multi-Class MRI Classification},
  journal = {i-manager's Journal on Image Processing},
  year = {2026},
  volume = {13},
  number = {2},
  pages = {1},
  doi = {10.26634/jip.13.2.11329}
}
Kostenlos starten →