Project
Prediction system for noisy brain medical image classification
Published
Designed and implemented a deep learning model from scratch to classify noisy brain medical images. The work explored custom CNN architectures and Vision Transformer variants for feature extraction and classification, with pipelines for noise handling, feature learning, and explainability using attention maps.
This project underpins the published ARiViT paper. The CV reports an 85% improvement in prediction accuracy for the implemented system.
Tech PyTorch · DataLoader · Pandas · Matplotlib · Scikit-learn · Jupyter Notebooks
Related research: ARiViT