Road Way to Safety
CNN Model
We are using YOLOv2 model architecture to detect potholes.
Model is trained on a dataset containing 259 images having a total of 829 potholes in them.
To detect potholes on any image or video just run the following command-
python predict.py -c config.json -w saved_weights.h5(path of the saved weight should be provided) -i (Image or video path)
