Sign in Subscribe
Artificial neural networks

Breakthrough in Brain Cancer Diagnosis: Ensemble-Based Convolutional Neural Networks

Strategic Insights

— 1 min read

Researchers at the University of Sevilla have developed an innovative approach to classify brain tumors using ensemble-based deep learning models. This method has shown significant promise in achieving high accuracy while maintaining interpretability for clinical use. By combining the strengths of multiple Convolutional Neural Network (CNN) architectures and incorporating explainability techniques, the team has created a more reliable and transparent diagnostic tool for medical professionals.

Key Takeaways:

  • The research used transfer learning with pre-trained architectures such as VGG16, DenseNet121, and Inception-ResNet-v2 to classify brain tumors from MRI images.
  • An ensemble-based classifier was developed using a majority voting strategy to improve robustness and achieve an accuracy of 86.17% in distinguishing gliomas, meningiomas, pituitary adenomas, and benign cases.
  • Interpretability techniques such as Grad-CAM++ and Integrated Gradients were employed, allowing visualization of model decision-making and identifying key regions influencing model predictions.
  • The proposed ensemble-based deep learning framework enhances the accuracy and interpretability of brain tumor classification from MRI images.
  • The research has been peer-reviewed and published in the journal Computers in Biology and Medicine.
  • The study was conducted by Luis Sanchez-Moreno and his team at the University of Sevilla's Robotics and Technology of Computers Lab.

Statistics:

  • The ensemble model achieved an accuracy of 86.17% in distinguishing gliomas, meningiomas, pituitary adenomas, and benign cases.
  • The proposed framework improved the accuracy of brain tumor classification from MRI images by achieving a higher accuracy rate compared to individual CNN architectures.
  • The research integrated explainability techniques such as Grad-CAM++ and Integrated Gradients to enhance the transparency and trustworthiness of the model's decision-making process.
  • The study utilized pre-trained architectures such as VGG16, DenseNet121, and Inception-ResNet-v2 to leverage their existing knowledge and improve classification accuracy.

Sources:

  • Ensemble-based Convolutional Neural Networks for brain tumor classification in MRI: Enhancing accuracy and interpretability using explainable AI. Computers in Biology and Medicine, 2025;195:110555.
  • NewsRx. University of Sevilla Reports Findings in Brain Cancer (Ensemble-based Convolutional Neural Networks for brain tumor classification in MRI: Enhancing accuracy and interpretability using explainable AI). Journal of Engineering. July 7, 2025; p 5971.

Read more