Document is current

Article History

Received: 18 June 2025

Accepted: 9 October 2025

First Online: 17 November 2025

Declarations

:

: The authors declare no competing interests.

: Factors like the size, intensity and shape of tumors make MRI interpretation challenging. Conventional techniques suffer from issues and limitations surrounding feature extraction and efficiency. Besides, deep learning architecture can outperform traditional techniques. The contributions of such works demonstrate that these models can save model time best classify tumors, but their computational time represented by cost limitations may be exponentially high. Also, several deep learning techniques still overfit datasets, and many do not generalize differently scanned images. Apparently, a deep learning framework that is computationally efficient and reasonably accurate for classifying specific MRI BTs should be developed by extracting hierarchical features. A lightweight and efficient deep learning framework method to supplement accuracy in MRI BT detection can potentially address many of the current limitations.

: This article does not contain directly any studies with human participants or animals performed by any of the authors. We used openly available datasets from kaggle.