Deep Learning for Invasive Ductal Carcinoma Detection

This repository contains the complete codebase for my research paper:

“Deep Learning Model for Invasive Ductal Carcinoma Detection in Histopathology Images”
Author: Dylan Jayabahu (Waterloo, Canada) Presented at: 2025 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)

This research introduces a deep learning pipeline for the detection of Invasive Ductal Carcinoma (IDC) in breast histopathology whole-slide images (WSIs), addressing both accuracy and interpretability for clinical decision-making.

---

Paper Highlights

• Conference Presentation: IEEE CCECE 2025
• Problem Addressed: Early and accurate detection of Invasive Ductal Carcinoma
• Approach: Deep learning with novel oversampling and heatmap visualization techniques
• Performance:
  • Balanced Accuracy: 89.06%
  • F1-Score: 86.68%
• Key Innovations:
  • Oversampling from biologically homogenous regions
  • High-resolution sliding window heatmaps

🔗 Full Paper PDF (Preprint): [https://doi.org/10.36227/techrxiv.172979260.09225931/v2]

🔗 Conference Program: [https://ccece2025.ieee.ca/technical-program/]