khanhha/crack_segmentation

This repository contains code and dataset for the task crack segmentation using two architectures UNet_VGG16, UNet_Resnet and DenseNet-Tiramusu

43
/ 100
Emerging

Implements semantic segmentation for pavement and concrete cracks using encoder-decoder UNet architectures with VGG16 and ResNet101 backbones for transfer learning. The solution aggregates 11,200 images from 12 public crack datasets into a unified 448×448 benchmark, trained with PyTorch and evaluated using IoU and Dice metrics. Addresses real-world robustness through diverse test scenarios including moss occlusion, texture noise, and large contextual backgrounds to reduce false positives.

431 stars. No commits in the last 6 months.

No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 8 / 25
Community 25 / 25

How are scores calculated?

Stars

431

Forks

138

Language

Python

License

Last pushed

May 06, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/khanhha/crack_segmentation"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.