Awesome

CARRADA-RAC: RA Annotation Calibration for CARRADA

This is the annotation calibration pipeline proposed in the paper PeakConv.

[PeakConv: Learning Peak Receptive Field for Radar Semantic Segmentation], CVPR 2023, accepted.

Pipeline:

Main Contributor for this code: Xinyan Zhang & Liwen Zhang

This repository groups the implemetations of CARRADA-RAC.

The code for this pipeline is based on the paper CARRADA

If you find this code useful for your research, please cite our paper:

@InProceedings{pkc2023,
	       author = {Liwen, Zhang and Xinyan, Zhang and Youcheng, Zhang and Yufei, Guo and Yuanpei, Chen and Xuhui, Huang and Zhe, Ma},
	       title = {PeakConv: Learning Peak Receptive Field for Radar Semantic Segmentation},
	       booktitle = {CVPR 2023 accepted},
	       month = {March},
	       year = {2023}
	       }

The Illustration of CARRADA-RAC annotation:

Download CARRADA dataset

Note: The CARRADA dataset used for this pipeline is considered as already downloaded by default. If it is not the case, you can follow the guidelines of the dedicated carrada_dataset repository.

Running the code

1. Run the carrada-rac's entire pipeline: It is the basic pipeline from carrada, you can skip and directlly skip to step two; It is mandatory to specify the path where the CARRADA dataset is located. Example: I put the Carrada folder in /datasets/, the path I should specify is /datasets/. If you are using Docker, the CARRADA dataset is extracted in the /datasets/ folder by default.

$ cd scripts/
$ bash run_annotation_pipeline.sh /datasets/

2. Running the first block to get new rd points:

$ cd scripts/
$ python generate_rd_points.py
$ python generate_centroids.py
$ python generate_annotation_files.py

3. Running the second block to get best ra points: It is the key step to get the new annotations.

$ cd scripts/
$ python get_best_ra.py
$ python get_new_annotation.py

Note:if you want to get the old(carrada) annotations:

$ cd scripts/
$ python get_old_annotation.py

Acknowledgements

• The Soft Dice loss is based on the code from https://github.com/kornia/kornia/blob/master/kornia/losses/dice.py
• Thank CARRADA dataset for providing the Radar dataset.
• Thank MVRSS for providing the basic model framework of multi-view RSS network. And we build our model using the basic framework of mvrss lib, our incremental modifications to the mvrss code did not effect the TMVA-Net and MV-Net in MVRSS.
• The paper is accepted by CVPR 2023. The camera-ready paper for PKC is coming soon, and we will provide the long extension version on arxiv.

License

The repo is released under the Apache 2.0 license.