The Strategy of Generalization Ability Improvement for Brain Tumor Classification Based on CNNs Model

DOI

10.2991/978-94-6463-540-9_63How to use a DOI?

Keywords

CNNs; generalization ability; brain tumor

Abstract

Brain tumor is a serious disease that affects lots of people. Traditional methods of tumor detection are time-consuming and subjective. Many studies have demonstrated Convolutional Neural Networks (CNNs) can classify brain tumors with a high accuracy, but they did not focus on the generalization of the model. This study proposes a way to enhance the generalization ability of CNN model for brain tumor classification. Two distinct sets were used in the study, one was designed to be the original dataset, the other was the external dataset. The testing set from the external dataset was split into different proportions (0%, 10%, 20%, 30%, 40%) then incorporated into the training set of original datasets to form 5 different training and testing sets. Five models with the same architecture were trained based on these training sets. The validation set used for training was from the original dataset in order to keep models align with the original distribution. Models were being tested on both the testing set of original datasets and the testing set been split. The 30% model turns out to have the best balance, which indicates that by incorporating a proper amount of external data into the training set, the model’s generalization ability could improve with a tradeoff of some accuracy.

Copyright

© 2024 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yuze Hou
PY  - 2024
DA  - 2024/10/16
TI  - The Strategy of Generalization Ability Improvement for Brain Tumor Classification Based on CNNs Model
BT  - Proceedings of the 2024 2nd International Conference on Image, Algorithms and Artificial Intelligence (ICIAAI 2024)
PB  - Atlantis Press
SP  - 628
EP  - 633
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-540-9_63
DO  - 10.2991/978-94-6463-540-9_63
ID  - Hou2024
ER  -