Computer Vision Classification of Breast Cancer Tumors via Deep Learning
I applied transfer learning with ResNet152 (PyTorch) to classify breast cancer tumors. The workflow includes data pre-processing, augmentations, and fine-tuning a pre-trained ResNet152 for both binary and multi-class classification.
Key results
- Binary classification (benign vs malignant) — Test accuracy: 99.3%, Test loss: 0.02 (ResNet152 transfer learning).
- Multi-class classification (6 tumor types) — Accuracy: 90.30% via transfer learning with ResNet152.
Artifacts
