Enhancing Prediction of Brain Tumor Classification Using Images and Numerical Data Features
Oumaima Saidani1, Turki Aljrees2, Muhammad Umer3
1Department of Information Systems, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman University, Riyadh 11671, Saudi Arabia.
Diagnostics (Basel, Switzerland)
|August 12, 2023
Summary
This study introduces advanced AI models for early brain tumor detection using image and data features. Both models achieved 0.99 accuracy, significantly improving diagnostic capabilities for neurological diseases.
Area of Science:
- Medical imaging analysis
- Artificial intelligence in healthcare
- Neurological disease diagnostics
Background:
- Brain tumors are a major cause of global mortality, necessitating early diagnosis for effective treatment.
- Accurate differentiation between individuals with and without tumors is critical for patient outcomes.
Purpose of the Study:
- To develop and validate AI models for accurate brain tumor classification using combined imaging and data features.
- To enhance early detection rates for neurological diseases, specifically brain tumors.
Main Methods:
- Image dataset enhancement followed by UNet transfer learning for tumor classification.
- Utilization of a voting classifier with 13 features, incorporating deep convolutional layers and combined gradient descent with logistic regression.
Main Results:
- Achieved a superior classification performance with an accuracy score of 0.99 for both proposed models.
- Demonstrated robust performance through comparison with existing supervised learning algorithms and state-of-the-art models.
Conclusions:
- The developed AI models show high efficacy in accurately diagnosing brain tumors.
- This approach offers a promising tool for early detection and improved management of neurological diseases.
