Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{194711,
        author = {B. Mukeswar Reddy and S. Aswani and B. Jayaprakash Narayana and V. Karthik Reddy and K. Manjunath},
        title = {AN EFFICIENT DEEP LEARNING FRAMEWORK FOR BRAIN TUMOR DETECTION USING MRI IMAGES},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {10},
        pages = {5250-5255},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=194711},
        abstract = {Scans help spot tumors in the brain. When detection happens fast, people live longer and get better care, so MRI plays a key role here. Bigger or smaller, shaped differently, sitting in various spots inside the skull - these make it hard for doctors to judge without help. Because experts must review each image one by one, mistakes can slip through; consistency often suffers under such pressure. Fast increases in medical image data mean automated tools now help doctors choose better paths forward. Thanks to advances in machine learning, particularly deep learning methods, smart systems now manage to understand complex scans well. One key player here is the CNN - a type of network built for image tasks - with its efficient net design standing out clearly.
Starting from existing weight values, the system first learns basic patterns. Then, only specific parts of the network adjust further, shaping itself to capture brain scan details more accurately. To handle variation and boost reliability, images are altered and scaled uniformly during learning. Instead of relying solely on precision metrics, outcomes are checked through loss trends, misclassification maps, and detailed performance tables. From the data, clear patterns emerge - convergence is steady, classification stays even, accuracy holds up across both cancer and normal types. Performance lines cluster tightly, indicating the method works well when spotting brain tumors. This setup appears robust enough for daily practice, offering helpful insights without adding confusion. Workers describe it as practical, trustworthy, built to handle actual cases. Radiologists may rely on these outputs when making patient judgments.},
        keywords = {},
        month = {March},
        }