CODE RELEASE

[wanghangege]

Great work, do you have any plans to open source the code?

[HHuiwen]

+1

[ldkong1205]

Hi @wanghangege and @HHuiwen, the code and dataset were released a while ago. Which part of the implementation are you looking for? @AlanLiangC, please help them with the code and data prepartion, thanks!

[HHuiwen]

@AlanLiangC
I noticed that the following two items in the to-do list section are not checked. I’m not sure whether the repository already includes the following:
• Release the whole Pi3DET dataset.
• Release the code for the Pi3DET-Net method.

[AlanLiangC]

Hello, sorry fot late update, the dataset is uploading to Huggingface now. But it's a little huge, maybe you can have fun with our competition version firstly. And our baseline models and Pi3DET-Net are already and will be updated at the competition codabase. Thanks for your interests again.

[liuzengyun]

Hello, @AlanLiangC , we note that you have already made the dataset available at huggingface. May we ask if this is the full dataset described in the ICCV25 paper? Does it include annotation information for the drone platforms? We are interested in conducting research based on your dataset and would greatly appreciate your response.

[AlanLiangC]

Hello, @AlanLiangC , we note that you have already made the dataset available at huggingface. May we ask if this is the full dataset described in the ICCV25 paper? Does it include annotation information for the drone platforms? We are interested in conducting research based on your dataset and would greatly appreciate your response.

Hi and yes, the dataset is the same as described in the paper. And for convenience, I will push a tookit for visiualization in one week. Thanks for your interest again.

[AlanLiangC]

Hello, @AlanLiangC , we note that you have already made the dataset available at huggingface. May we ask if this is the full dataset described in the ICCV25 paper? Does it include annotation information for the drone platforms? We are interested in conducting research based on your dataset and would greatly appreciate your response.

Hi and yes, the dataset is the same as described in the paper. And for convenience, I will push a tookit for visiualization in one week. Thanks for your interest again.

We have publised a visiualization toolkit for Pi3DET-Dataset. Have fun at Pi3DET-Visualization

[liuzengyun]

Hello, @AlanLiangC , we note that you have already made the dataset available at huggingface. May we ask if this is the full dataset described in the ICCV25 paper? Does it include annotation information for the drone platforms? We are interested in conducting research based on your dataset and would greatly appreciate your response.

Hi and yes, the dataset is the same as described in the paper. And for convenience, I will push a tookit for visiualization in one week. Thanks for your interest again.

@AlanLiangC Thank you very much for your reply. I would like to access the annotations for the drone platform data (with 5885 frames in robosense-track5). May I ask if they are located in the processed/Drone/Outdoor_Day/***/ps_labels_final/final_ps_dict.pkl file?
Or how should I proceed to obtain a pkl file with annotations similar to those on the Car platform(https://huggingface.co/datasets/robosense/datasets/blob/main/track5-cross-platform-3d-object-detection/phase12_vehicle_training/pi3det_infos_train.pkl)?

[AlanLiangC]

Hello, @AlanLiangC , we note that you have already made the dataset available at huggingface. May we ask if this is the full dataset described in the ICCV25 paper? Does it include annotation information for the drone platforms? We are interested in conducting research based on your dataset and would greatly appreciate your response.

Hi and yes, the dataset is the same as described in the paper. And for convenience, I will push a tookit for visiualization in one week. Thanks for your interest again.

@AlanLiangC Thank you very much for your reply. I would like to access the annotations for the drone platform data (with 5885 frames in robosense-track5). May I ask if they are located in the processed/Drone/Outdoor_Day/***/ps_labels_final/final_ps_dict.pkl file? Or how should I proceed to obtain a pkl file with annotations similar to those on the Car platform(https://huggingface.co/datasets/robosense/datasets/blob/main/track5-cross-platform-3d-object-detection/phase12_vehicle_training/pi3det_infos_train.pkl)?

Annotations used in robosense-track5 is a subset of the whole Drone-Platform-Data, as it is a split for an official competition, I must request the organizer's confirmation before realeasing it. But you can make breakpoints within visiualization tookit or check the class method to learn how to extrack the whole data and annotations.