A Comprehensive Review of Brain Tumor Using Deep Learning
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
The ability to identify brain tumors using magnetic resonance imaging (MRI) has become an important aspect in enhancing the accuracy of detection procedures as well as the treatment planning. Conventional machine learning algorithms can usually be quite sensitive to changes in the shape, intensity, and texture of tumors; hence, deep learning has developed into a groundbreaking remedy. As recent research, such as points out, much more impressive results on tumor localization and classification are provided by convolutional neural networks (CNNs), U-Net versions, variants equipped with GAN, and transformer-based models. Embodied within larger models, U-Net-based GAN models and the DacFormer generative adversarial network shown to be better at preserving boundaries and resisting noise, and multimodal Transformerbased models such as TransBTS are shown to be less prone to boundary failures. In sum, recent developments in deep learning innovations in the sphere of brain tumor diagnosis and segmentation that can be identified between 2018 and 2025 will be reviewed through (1) highlighting progress in the fields of CNN, U-Net, GAN, and Transformer as well as ensemblebased methods; (2) evaluating the dynamics of AI performance in terms of accuracy and sensitivity across datasets, architectures, and optimization algorithms; (3) describing the contributions to the discussion using explainable and clinically deployable AI solutions; and (4) revealing the limitations of this research.
Fields of Study & Subjects
Brain Tumor Detection and Classification
Medical Image Segmentation Techniques
AI in cancer detection
Publication details
- Journal / source
- International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT)
- Pages
- 956–963
- Publication date
- 4 February 2026
- Language
- EN