DCAITI Project--WS20/21
1.You can train your own version or download the weight files from https://drive.google.com/drive/folders/1gFtHe5ZCGfZnrlUW-Pua1UCJY1JM6ZFR?usp=sharing and put the .pth files into the corresponding folder.
2.For the develop of rcnn, you need to set the fasterrcnn_resnet50_fpn_coco.pth to ./RCD_FRCNN/backbone.
3.For the deployment on android devices, please see ./RCD_nanodet/demo_android_ncnn/README.md.
4.Download the dataset: https://github.com/sekilab/RoadDamageDetector
Reference:
Nanodet:https://github.com/RangiLyu/nanodet.git
EfficientDet:https://github.com/mahdi65/roadDamageDetection2020