Publication Type : Conference Paper
Publisher : IEEE
Source : 2023 3rd International Conference on Intelligent Technologies (CONIT)
Url : https://doi.org/10.1109/conit59222.2023.10205710
Campus : Kochi
School : School of Computing
Year : 2023
Abstract : One of the main reasons for mortality worldwide is cancer. There are different types of cancers, among them, brain tumor patients have a low survival rate. Brain tumors are of different categories and are mainly differentiated based on its size and where it is present. Since brain tumors are severe, timely detection is vital. This brings in the necessity for a computer-assisted System, which helps doctors classify brain tumors into their different types and treat them accordingly. So here, we put forward a ResNet-50 model along with dimensionality reduction and feature selection techniques to categorize the MR-Images into 4 main kinds-meningioma, glioma, pituitary, and no-tumor. The suggested approach has obtained the highest accuracy of 98.6%, in comparison with other models using the same dataset.
Cite this Research Publication : Anjana M Nair, Lakshmi S Kumar, Vimina E R, Multiclass Brain Tumor Classification of MR-Images Using ResNet-50 and Dimensionality Reduction Techniques, 2023 3rd International Conference on Intelligent Technologies (CONIT), IEEE, 2023, https://doi.org/10.1109/conit59222.2023.10205710