Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Oct 7:10:20552076241286140.
doi: 10.1177/20552076241286140. eCollection 2024 Jan-Dec.

Brain tumor classification using fine-tuned transfer learning models on magnetic resonance imaging (MRI) images

Sadia Maduri Rasa  1 , Mohammed Manowarul Islam  1 , Mohammed Alamin Talukder  2 , Mohammed Ashraf Uddin  3 , Majdi Khalid  4 , Mohsin Kazi  5 , Mohammed Zobayer Kazi  1
Affiliations

Brain tumor classification using fine-tuned transfer learning models on magnetic resonance imaging (MRI) images

Sadia Maduri Rasa et al. Digit Health. .

Abstract

Objective: Brain tumors are a leading global cause of mortality, often leading to reduced life expectancy and challenging recovery. Early detection significantly improves survival rates. This paper introduces an efficient deep learning model to expedite brain tumor detection through timely and accurate identification using magnetic resonance imaging images.

Methods: Our approach leverages deep transfer learning with six transfer learning algorithms: VGG16, ResNet50, MobileNetV2, DenseNet201, EfficientNetB3, and InceptionV3. We optimize data preprocessing, upsample data through augmentation, and train the models using two optimizers: Adam and AdaMax. We perform three experiments with binary and multi-class datasets, fine-tuning parameters to reduce overfitting. Model effectiveness is analyzed using various performance scores with and without cross-validation.

Results: With smaller datasets, the models achieve 100% accuracy in both training and testing without cross-validation. After applying cross-validation, the framework records an outstanding accuracy of 99.96% with a receiver operating characteristic of 100% on average across five tests. For larger datasets, accuracy ranges from 96.34% to 98.20% across different models. The methodology also demonstrates a small computation time, contributing to its reliability and speed.

Conclusion: The study establishes a new standard for brain tumor classification, surpassing existing methods in accuracy and efficiency. Our deep learning approach, incorporating advanced transfer learning algorithms and optimized data processing, provides a robust and rapid solution for brain tumor detection.

Keywords: AdaMax optimizer.; Adam; Brain tumor detection; convolutional neural network; deep learning; magnetic resonance imaging; transfer learning.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Overview diagram of the proposed methodology.
Figure 2.
Figure 2.
Few samples of magnetic resonance imaging (MRI) images (a) before and (b) after pre-processing.
Figure 3.
Figure 3.
Samples of augmented brain magnetic resonance imaging (MRI) images.
Figure 4.
Figure 4.
Main two blocks of MobileNetV2.
Figure 5.
Figure 5.
EfficientNetB3 architecture of our proposed methodology: (a) The original EfficientNetB3 model and (b) fine-tuned EfficientNetB3 model.
Figure 6.
Figure 6.
Performance results of MobileNetV2 using Adam optimizer for the two-class dataset: (a) actual versus validation loss; (b) actual versus validation accuracy; (c) confusion matrix before normalization; and (d) confusion matrix after normalization.
Figure 7.
Figure 7.
Performance results of MobileNetV2 using AdaMax optimizer for the two-class dataset: (a) actual versus validation loss; (b) actual versus validation accuracy; (c) confusion matrix before normalization; and (d) confusion matrix after normalization.
Figure 8.
Figure 8.
Performance results of MobileNetV2 using Adam optimizer for the four-class dataset: (a) actual versus validation loss; (b) actual versus validation accuracy; (c) confusion matrix before normalization; and (d) confusion matrix after normalization.
Figure 9.
Figure 9.
Performance results of MobileNetV2 using AdaMax optimizer for the four-class dataset: (a) actual versus validation loss; (b) actual versus validation accuracy; (c) confusion matrix before normalization; and (d) confusion matrix after normalization.
Figure 10.
Figure 10.
Some samples of misclassified images.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Park KS. Nervous system. In: Humans and electricity: Understanding body electricity and applications. Korea: Springer Cham, 2023, pp.27–51.
    1. Talukder MA, Islam MM, Uddin MA, et al.. An efficient deep learning model to categorize brain tumor using reconstruction and fine-tuning. Expert Syst Appl 2023; 230: 120534.
    1. Evans-Martin F. The nervous system. New York: Infobase Holdings, Inc., 2022.
    1. Legler JM, Ries LAG, Smith MA, et al.. Brain and other central nervous system cancers: recent trends in incidence and mortality. J Nat Cancer Inst 1999; 91: 1382–1390. - PubMed
    1. Packer RJ, Gurney JG, Punyko JA, et al.. Long-term neurologic and neurosensory sequelae in adult survivors of a childhood brain tumor: childhood cancer survivor study. J Clin Oncol 2003; 21: 3255–3261. - PubMed

LinkOut - more resources