Please post links to your external data sources here before the deadline specified (March 28, 2023). This includes datasets and pre-trained models. Once it has been posted, you do not need to post it again.
MAFAT Challenge Team
Posted by: MAFAT_Challenge @ Feb. 2, 2023, 7:02 a.m.Dear organizers, I'd like to use following datasets in this competition. Can you please approve whether we are allowed to use these in competition:
1. DOTA: https://captain-whu.github.io/DOTA/dataset.html
2. HRSC2016: https://www.kaggle.com/datasets/guofeng/hrsc2016
3. UCAS-AOD: https://opendatalab.com/102/download
4. Dior Dataset Image Dataset: https://universe.roboflow.com/new-workspace-ghppr/dior-dataset-riv6b/dataset/2
5. DIUx xView 2018 Detection Challenge: http://xviewdataset.org/
Thanks!
Posted by: BloodAxe @ March 6, 2023, 7:50 p.m.Hi,
Please note that it is the participant’s responsibility to ensure that the dataset complies with the conditions of the competition and that there is no violation of the dataset license.
From a quick and non-obligating review - we think there will not be a problem using the datasets described in sections: 1,2,4.
Regarding UCAS-AOD and xView - we could not find a clear definition of the dataset license.
MAFAT Challenge Team
We would like to use the pretrained model in https://github.com/ViTAE-Transformer/Remote-Sensing-RVSA.
Posted by: wangzhiyu918 @ March 26, 2023, 2:51 a.m.Hi,
From a quick and non-obligating review - we think there will not be a problem using the pretrained models from that link.
Please note that it is the participant’s responsibility to ensure that the external data used complies with the conditions of the competition and that there is no violation of the dataset license.
MAFAT Challenge Team
Pretrained models:
1. pretrained weight in timm github repo: https://github.com/huggingface/pytorch-image-models
2. pretrained weight in mmrotate github repo: https://github.com/open-mmlab/mmrotate
3. yolov5 weight: https://github.com/ultralytics/yolov5
4. pretrained models in OBBDetection github repo: https://github.com/jbwang1997/OBBDetection
External dataset:
1. DOTA: https://captain-whu.github.io/DOTA/dataset.html
2. HRSC2016: https://www.kaggle.com/datasets/guofeng/hrsc2016
Dear MAFAT Challenge Team,
We would like to use the following:
Public Dataset:
https://captain-whu.github.io/DOTA/dataset.html
Pre-trained models:
https://github.com/hukaixuan19970627/yolov5_obb
Hi,
I would like to use this dataset:
- SODA-A: https://shaunyuan22.github.io/SODA/ (license: MIT, reference: https://github.com/shaunyuan22/SODA/blob/main/LICENSE)
Should we specify pretrained models used on mmrotate? Or the one stated by nvnn is more than enough?
Posted by: rasyidridha @ March 27, 2023, 1:35 p.m.Pretrained models:
https://github.com/ViTAE-Transformer/Remote-Sensing-RVSA
https://github.com/zcablii/Large-Selective-Kernel-Network
https://github.com/jbwang1997/OBBDetection
https://github.com/hukaixuan19970627/yolov5_obb
https://github.com/NVIDIA/retinanet-examples
https://github.com/hukaixuan19970627/yolov5_obb
https://github.com/BossZard/rotation-yolov5
https://github.com/jbwang1997/OBBDetection
https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/rotate
https://github.com/DDGRCF/YOLOX_OBB
Public Datasets:
Cofga - the dataset from the old mafat competition - https://github.com/bok11/02456-MAFAT/tree/master/dataset_v2/root
visdrone - https://github.com/VisDrone/VisDrone-Dataset
COWC - https://gdo152.llnl.gov/cowc/
iSAID - https://captain-whu.github.io/iSAID/dataset.html
xView - http://xviewdataset.org/
xView2 - https://xview2.org/dataset
S2Looking - https://github.com/S2Looking/Dataset
HRSC2016 - https://www.kaggle.com/datasets/guofeng/hrsc2016
I would like to also get permission to use any publicly avaliable dataset on kaggle
Posted by: OFire @ March 28, 2023, 8:59 p.m.Hi,
Each dataset must be declared specifically, therefore a "bundle" declaration such as "all of publicly available datasets on kaggle" is not valid.
Also, please note that it is the participants responsibility to make sure the datasets they use are inline with the competition rules.
Shai,
MAFAT Challenge Team