Optimizing Tumor Detection in Brain MRI with Machine Learning Algorithms

Strategic Insights -

Researchers at the University of Valencia have made significant breakthroughs in the detection of brain tumors using machine learning algorithms, particularly One-Class Support Vector Machines (OCSVM) and Convolutional Neural Networks (CNN). The study sought to address the class imbalance issue in medical imaging datasets, which poses significant challenges for traditional classification algorithms. By employing a hybrid approach that combines CNN-based feature extraction with OCSVM, the researchers demonstrated significant improvements in anomaly detection performance.

Key Takeaways:

  • The researchers used a One-Class Support Vector Machine (OCSVM) trained exclusively on features extracted from healthy brain MRI images, using deep learning architectures such as DenseNet121, VGG16, MobileNetV2, InceptionV3, and ResNet50.
  • Experimental results showed that combining Convolutional Neural Network (CNN)-based feature extraction with OCSVM significantly improved anomaly detection performance compared to simpler handcrafted approaches.
  • DenseNet121 achieved an accuracy of 94.83%, a precision of 99.23%, and a sensitivity of 89.97%, while VGG16 reached an accuracy of 95.33%, a precision of 98.87%, and a sensitivity of 91.32%.
  • MobileNetV2 showed a competitive trade-off between accuracy (92.83%) and computational efficiency, making it suitable for resource-constrained environments.
  • The pure CNN model-trained directly for classification without OCSVM-outperformed hybrid methods with an accuracy of 97.83%, highlighting the effectiveness of deep convolutional networks in directly learning discriminative features from MRI data.
  • This approach enables reliable detection of brain tumor anomalies without requiring labeled pathological data, offering a promising solution for clinical contexts where abnormal samples are scarce.
  • Future research will focus on reducing inference time, expanding and diversifying training datasets, and incorporating explainability tools to support clinical integration and trust in AI-based diagnostics.
  • The study was led by Alfredo Rosado-Munoz, GDDP, Dept. of Electronic Engineering, School of Engineering, University of Valencia, 46100 Burjassot, Valencia, Spain.

Statistics:

  • DenseNet121 achieved an accuracy of 94.83%, a precision of 99.23%, and a sensitivity of 89.97%.
  • VGG16 reached an accuracy of 95.33%, a precision of 98.87%, and a sensitivity of 91.32%.
  • MobileNetV2 showed an accuracy of 92.83%, a precision of 97.45%, and a sensitivity of 87.95%.
  • The pure CNN model-trained directly for classification without OCSVM-outperformed hybrid methods with an accuracy of 97.83%.

Sources:

  • NewsRx. University of Valencia Reports Findings in Support Vector Machines (Optimizing Tumor Detection in Brain MRI with One-Class SVM and Convolutional Neural Network-Based Feature Extraction). Journal of Engineering. August 4, 2025; p 5435.
  • Journal of Imaging, 2025;11(7):207. (Optimizing Tumor Detection in Brain MRI with One-Class SVM and Convolutional Neural Network-Based Feature Extraction)