Domain Adaptive Faster R-CNN for Object Detection in the Wild
This is the implementation of our CVPR 2018 work 'Domain Adaptive Faster R-CNN for Object Detection in the Wild'. The aim is to improve the cross-domain robustness of object detection, in the screnario where training and test data are drawn from different distributions. The original paper can be found here.
If you find it helpful for your research, please consider citing:
@inproceedings{chen2018domain,
title={Domain Adaptive Faster R-CNN for Object Detection in the Wild},
author={Chen, Yuhua and Li, Wen and Sakaridis, Christos and Dai, Dengxin and Van Gool, Luc},
booktitle = {Computer Vision and Pattern Recognition (CVPR)},
year={2018}
}
If you encounter any problems with the code, please contact me at yuhua[dot]chen[at]vision[dot]ee[dot]ethz[dot]ch
An example of adapting from Cityscapes dataset to Foggy Cityscapes dataset is provided:
Download the datasets from here. Specifically, we will use gtFine_trainvaltest.zip, leftImg8bit_trainvaltest.zip and leftImg8bit_trainvaltest_foggy.zip.
Prepare the data using the scripts in 'prepare_data/prepare_data.m'.