| Autor: |
Priyadharshini, K., Dhamodaran, M., Kaviyapriya, S., Harini, R., Kaviya, V., Ramasubramanian, B. |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 2023, Vol. 2914 Issue 1, p1-6, 6p |
| Abstrakt: |
Brain tumors are the most common and antagonistic disease. As a result, its duration is limited. In this way, exact brain tumor distinguishing proof is fundamental for conceiving a genuine therapy methodology to fix and expand the existence of individuals with cerebrum tumors. The goal of data augmentation and fine-tuning methods is to separate and classify tumors. In this paper, we perform brain tumor classification in the proposed framework using Jupiyter Notebook using four different types of brain tumor datasets obtained from Kaggle. The proposed work employs three different Convolutional Neural Networks (CNN) architectures, namely Alex Net, Google Net, and VGG-16Net, on MRI slices to identify tumor types. It is observed that the proposed method achieved an accuracy of 98.74% for VGG-16 Net which is comparatively high than the state of art tumor detection method. The developed method is also compared with two other transfer learning networks such as Alex Net and Google Net. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
