Reproduce “3D Brain MRI Classification for Alzheimer Diagnosis Using CNN with Data Augmentation” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Thien Nhan Vo, Bac Nam Ho

A study applied a 3D Convolutional Neural Network for binary classification of Alzheimer's disease from brain MRI scans, investigating the impact of horizontal flipping data augmentation. The models exhibited severe overfitting and failed to generalize due to an extremely small dataset and a labeling issue in the test set.

Key insights

  • Even simple data augmentation like horizontal flipping is insufficient to prevent severe overfitting when the original training dataset is extremely small (18 samples).
  • Rigorous data preprocessing and validation are critical, as a test set labeling error can invalidate all experimental results.
  • Complex deep learning models like 3D CNNs require substantially larger and diverse datasets for robust feature learning and generalization in medical imaging.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.