R3AD is an active dataset that uses a dense sampling scheme, as shown in the following figure:
Here are some examples data:
There are seven scene in our dataset. They can be downloaded at:
Home_1: https://drive.google.com/file/d/1-P4l5Sicu_inXMP_w8xoJnZL956Ms50W/view?usp=sharing
Home_2: https://drive.google.com/file/d/1-1521PLjtWaYETOjHzmv5k40brDuRsW0/view?usp=sharing
Home_3: https://drive.google.com/file/d/1-LLcHuRJE1qcSXwzyYaveuiqYM-qRHX1/view?usp=sharing
Home_4: https://drive.google.com/file/d/1IDG6epKH63dIQAE08dRY-3jyWUH39OPf/view?usp=sharing
Home_5: https://drive.google.com/file/d/1-0tgBw4yWPKYrHsTnpAQRNoUVzFyclei/view?usp=sharing
Home_6: https://drive.google.com/file/d/1-02rVp1IaWLrAoRRvfLhFgvJOnG8d1V7/view?usp=sharing
Home_7: https://drive.google.com/file/d/13uZ-7NPlbu-meB_gmwDfJa40u4FGtJ2C/view?usp=sharing
Here are some demo code for loading data and annotation:
https://drive.google.com/file/d/1-aBXN7to9mcrUMrkLMjq6Mm06LeVP3s5/view?usp=sharing
Because we modified many APIs in mmdetection3d and tianshou, there may be some bugs in current version. We will continue to update the code in the future.
mmdetection3d: https://github.com/open-mmlab/mmdetection3d
tianshou: https://github.com/thu-ml/tianshou
This project is released under the MIT-license
Under review.