Publication Type : Conference Proceedings
Publisher : IEEE
Source : 2024 5th International Conference on Smart Electronics and Communication (ICOSEC), Trichy, India, 2024, pp. 1210-1216, doi: 10.1109/ICOSEC61587.2024.10722290. (28.10.2024-xplore)
Url : https://ieeexplore.ieee.org/document/10722290
Campus : Bengaluru
School : School of Engineering
Department : Electrical and Electronics
Year : 2024
Abstract : There has been increasing concern about the rising rates of brain tumors. Brain tumors can be classified as primary and secondary tumors. Primary tumors are developed in the brain and the secondary tumors are the one which spread to the brain from other parts of the body. Approximately 120 distinct types of brain tumors have been identified. Magnetic Resonance Imaging (MRI) scans are considered the most effective diagnostic tool for detecting breaches in the blood-brain barrier. The advantages of using MRI scanning for brain analysis include clear visualization of soft tissue without any side effects from ionizing radiation. Manual diagnosis by radiologists and physicians is time consuming, costly, and prone to inaccuracies. For effective treatment, timely detection of brain tumors is must. Computer-aided diagnosis using machine learning algorithms, along with digital image processing is effective in brain tumor classification. Convolutional Neural Network (CNN) is proposed in this paper for identifying the brain tumor because in CNN more features are extracted. Brain tumor images are collected from the patients MRI. CNN classifier is tested on these clinical data. Initially there was overfitting problem which was addressed using k-fold cross validation. The highest accuracy of the model was found to be 94.89%, precision 0.9448 with 3-fold cross validation for 5 dense layers.
Cite this Research Publication : Swapna Sanapala; M. R. Rashmi; Tolga özer, "Brain Tumor Identification using Convolutional Neural Network," 2024 5th International Conference on Smart Electronics and Communication (ICOSEC), Trichy, India, 2024, pp. 1210-1216, doi: 10.1109/ICOSEC61587.2024.10722290. (28.10.2024-xplore)