- Employed Roboflow for preparing the custom dataset and applying data preprocessing and augmentation techniques like flip, saturation, blur.
- Utilized YOLOv8 instance segmentation for real-time, accurate detection and segmentation of concrete cracks in building and sculpture images.
- Deployed the model in Roboflow and evaluated the model performance using real images different from dataset.
- I used Roboflow visualization of training resluts and graphs.
- I collected the dataset from the google images and some from the roboflow public datasets.
- TensorFlow
I have deployed this model in the Robloflow platform.
https://app.roboflow.com/tjiitpkd/concretecrackdetection/deploy/1
you can find the results from this link
https://app.roboflow.com/tjiitpkd/concretecrackdetection/deploy/1