• Hena Fatma
• Dr. Sunil
• Dr. Md. Masroor Ahmad

Brain tumours represent a critical medical challenge that requires precise detection and diagnosis, particularly in the context of magnetic resonance imaging (MRI). Traditional image processing methods and conventional machine learning techniques often struggle with accurately pinpointing tumour locations in complex MRI scans, which are often affected by noise and inconsistent image quality. The advent of artificial intelligence (AI) has revolutionised many areas of healthcare, offering new possibilities for diagnostic and therapeutic advancements. This study focuses on exploring how AI-based techniques can improve accuracy in the automated detection and segmentation of brain tumours. Conducted within the Python programming environment, the research utilises MRI datasets specifically designed for brain tumour detection. These datasets are available for use in compliance with regulations set forth by the US Department of Health. The study’s results show that the earlier model achieved a convergence rate of approximately 90% accuracy for both training and validation datasets. While this suggests the model could identify data patterns, it fell short of delivering precise results, likely due to issues such as insufficient model complexity and inappropriate hyperparameters, including learning rate or batch size. In contrast, the MobileNet model demonstrated a substantial improvement, reaching 95% accuracy in both training and validation, marking a clear advancement over the previous model. MobileNet’s deeper architecture and use of pre-trained features allow for better generalization, evident in its consistent improvement in accuracy without overfitting. While the earlier model may suffice for simpler tasks, MobileNet proves more effective for complex datasets requiring higher precision. The proposed AI-based approach, which incorporates enhanced image pre-processing, boosts diagnostic accuracy and efficiency, ultimately aiding healthcare professionals in improving patient outcomes.

Accuracy, AI, Automated Detection, Segmentation, Brain Tumours, AI-driven Methodology.

"Enhancing Accuracy with AI-Powered Methods for Automated Detection and Segmentation of Brain Tumours", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 1, page no.g736-g742, January-2025, Available :http://www.jetir.org/papers/JETIR2501678.pdf

"Enhancing Accuracy with AI-Powered Methods for Automated Detection and Segmentation of Brain Tumours", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 1, page no. ppg736-g742, January-2025, Available at : http://www.jetir.org/papers/JETIR2501678.pdf