Brain Tumor Diagnosis Using Machine Learning, Convolutional Neural Networks, Capsule Neural Networks and Vision...

Area of Science:

• Neuro-oncology
• Artificial Intelligence in Medicine
• Medical Imaging Analysis

Background:

• Brain tumor grading is crucial for treatment planning, relying on clinical and radiological data.
• Current non-invasive methods for tumor grading are essential for optimal treatment selection.
• Deep Learning (DL) techniques, particularly Convolutional Neural Networks (CNNs), show promise in brain tumor diagnosis but have limitations.

Purpose of the Study:

• To provide a comprehensive overview of machine learning (ML)-based techniques for brain tumor classification and segmentation.
• To focus on the advancements and applications of CNNs, Capsule Neural Networks (CapsNets), and Vision Transformers (ViTs) in neuro-oncology.
• To discuss current challenges, limitations, and future research directions in DL for brain tumor analysis.

Main Methods:

• Review of recent literature on ML, DL, CNNs, CapsNets, and ViTs for brain tumor classification and segmentation.
• Analysis of the fundamental contributions and performance of state-of-the-art techniques.
• In-depth discussion of critical issues, open challenges, and limitations in the field.

Main Results:

• CNNs are effective for brain tumor diagnosis but struggle with input variations.
• CapsNets offer improved resistance to rotations and translations, beneficial for medical imaging.
• ViT-based solutions address long-range dependencies, enhancing CNN performance.
• The survey highlights the performance of various ML-based techniques in recent studies.

Conclusions:

• Advanced DL architectures like CapsNets and ViTs show significant potential for non-invasive brain tumor grading and segmentation.
• Addressing limitations in current DL models is key for future advancements in neuro-oncology.
• This survey serves as a foundation for further research into AI-driven brain tumor management.