🏗️ U-Net Based Crack Severity Classification

📌 Note: If you're unable to view the embedded Kaggle notebook, you can open it directly using this link:
🔗 Kaggle Notebook: Crack Segmentation using U-Net

---

A deep learning solution for automatic crack detection and severity assessment in structural images using modified U-Net architecture.

---

📌 Project Overview

Multi-task learning model that performs:

1. Pixel-wise segmentation of cracks
2. Severity classification into:
   • Class 0: No crack ✅
   • Class 1: Minor (<1mm) 🟢
   • Class 2: Moderate (1–3mm) 🟡
   • Class 3: Severe (>3mm) 🔴

---

📽️ Demo Video

---

🔗 Kaggle Notebook

Explore the full model training, segmentation results, and classification pipeline here:
📘 Crack Segmentation using U-Net (Kaggle Notebook)

---