CVPR2023-Occupancy-Prediction-Challenge
The world's First 3D Occupancy Benchmark for Scene Perception in Autonomous Driving.
Understanding the 3D surroundings including the background stuffs and foreground objects is important for autonomous driving. In the traditional 3D object detection task, a foreground object is represented by the 3D bounding box. However, the geometrical shape of the object is complex, which can not be represented by a simple 3D box, and the perception of the background stuffs is absent. The goal of this task is to predict the 3D occupancy of the scene. In this task, we provide a large-scale occupancy benchmark based on the nuScenes dataset. The benchmark is a voxelized representation of the 3D space, and the occupancy state and semantics of the voxel in 3D space are jointly estimated in this task. The complexity of this task lies in the dense prediction of 3D space given the surround-view images.
If you use the challenge dataset in your paper, please consider citing OccNet and Occ3D with the following BibTex:
@article{sima2023_occnet,
title={Scene as Occupancy},
author=author={Chonghao Sima and Wenwen Tong and Tai Wang and Li Chen and Silei Wu and Hanming Deng and Yi Gu and Lewei Lu and Ping Luo and Dahua Lin and Hongyang Li},
year={2023},
eprint={2306.02851},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
@article{tian2023occ3d,
title={Occ3D: A Large-Scale 3D Occupancy Prediction Benchmark for Autonomous Driving},
author={Tian, Xiaoyu and Jiang, Tao and Yun, Longfei and Wang, Yue and Wang, Yilun and Zhao, Hang},
journal={arXiv preprint arXiv:2304.14365},
year={2023}
}
active
)Please refer to this link. If you wish to add new / modify results to the leaderboard, please drop us an email to [email protected]
Top 10 at a glance by June 10 2023.
Given images from multiple cameras, the goal is to predict the current occupancy state and semantics of each voxel grid in the scene. The voxel state is predicted to be either free or occupied. If a voxel is occupied, its semantic class needs to be predicted, as well. Besides, we also provide a binary observed/unobserved mask for each frame. An observed voxel is defined as an invisible grid in the current camera observation, which is ignored in the evaluation stage.
Leaderboard ranking for this challenge is by the intersection-over-union (mIoU) over all classes.
Let $C$ be he number of classes.
$$ mIoU=\frac{1}{C}\displaystyle \sum_{c=1}^{C}\frac{TP_c}{TP_c+FP_c+FN_c}, $$
where $TP_c$ , $FP_c$ , and $FN_c$ correspond to the number of true positive, false positive, and false negative predictions for class $c_i$.
Type | Info |
---|---|
mini | 404 |
train | 28,130 |
val | 6,019 |
test | 6,008 |
cameras | 6 |
voxel size | 0.4m |
range | [-40m, -40m, -1m, 40m, 40m, 5.4m] |
volume size | [200, 200, 16] |
#classes | 0 - 17 |
The dataset contains 18 classes. The definition of classes from 0 to 16 is the same as the nuScenes-lidarseg dataset. The label 17 category represents voxels that are not occupied by anything, which is named as free
. Voxel semantics for each sample frame is given as [semantics]
in the labels.npz.
How are the labels annotated? The ground truth labels of occupancy derive from accumulative LiDAR scans with human annotations, and we annotate the occupancy in the ego coordinate system.
free
;[mask_lidar]
is a 0-1 binary mask, where 0's represent unobserved voxels. As shown in Fig.1(b), grey voxels are unobserved. Due to the limitation of the visualization tool, we only show unobserved voxels at the same height as the ground.Camera visibility. Note that the installation positions of LiDAR and cameras are different, therefore, some observed voxels in the LiDAR view are not seen by the cameras. Since we focus on a vision-centric task, we provide a binary voxel mask [mask_camera]
, indicating whether the voxels are observed or not in the current camera view. As shown in Fig.1(c), white voxels are observed in the accumulative LiDAR view but unobserved in the current camera view.
Both [mask_lidar]
and [mask_camera]
masks are optional for training. Participants do not need to predict the masks. Only [mask_camera]
is used for evaluation; the unobserved voxels are not involved during calculating the F-score and mIoU.
The files mentioned below can also be downloaded via OpenDataLab.It is recommended to use provided command line interface for acceleration.
Subset | Google Drive | Baidu Cloud | Size |
---|---|---|---|
mini | data | data | approx. 440M |
trainval | data | data | approx. 32G |
test | data | data | approx. 6G |
imgs
, gts
and annotations
. The imgs
datas have the same hierarchy with the image samples in the original nuScenes dataset.The hierarchy of folder Occpancy3D-nuScenes-V1.0/
is described below:
└── Occpancy3D-nuScenes-V1.0
|
├── mini
|
├── trainval
| ├── imgs
| | ├── CAM_BACK
| | | ├── n015-2018-07-18-11-07-57+0800__CAM_BACK__1531883530437525.jpg
| | | └── ...
| | ├── CAM_BACK_LEFT
| | | ├── n015-2018-07-18-11-07-57+0800__CAM_BACK_LEFT__1531883530447423.jpg
| | | └── ...
| | └── ...
| |
| ├── gts
| | ├── [scene_name]
| | | ├── [frame_token]
| | | | └── labels.npz
| | | └── ...
| | └── ...
| |
| └── annotations.json
|
└── test
├── imgs
└── annotations.json
imgs/
contains images captured by various cameras.gts/
contains the ground truth of each sample. [scene_name]
specifies a sequence of frames, and [frame_token]
specifies a single frame in a sequence.annotations.json
contains meta infos of the dataset.labels.npz
contains [semantics]
, [mask_lidar]
, and [mask_camera]
for each frame.annotations {
"train_split": ["scene-0001", ...], <list> -- training dataset split by scene_name
"val_split": list ["scene-0003", ...], <list> -- validation dataset split by scene_name
"scene_infos" { <dict> -- meta infos of the scenes
[scene_name]: { <str> -- name of the scene.
[frame_token]: { <str> -- samples in a scene, ordered by time
"timestamp": <str> -- timestamp (or token), unique by sample
"camera_sensor": { <dict> -- meta infos of the camera sensor
[cam_token]: { <str> -- token of the camera
"img_path": <str> -- corresponding image file path, *.jpg
"intrinsic": <float> [3, 3] -- intrinsic camera calibration
"extrinsic":{ <dict> -- extrinsic parameters of the camera
"translation": <float> [3] -- coordinate system origin in meters
"rotation": <float> [4] -- coordinate system orientation as quaternion
}
"ego_pose": { <dict> -- vehicle pose of the camera
"translation": <float> [3] -- coordinate system origin in meters
"rotation": <float> [4] -- coordinate system orientation as quaternion
}
},
...
},
"ego_pose": { <dict> -- vehicle pose
"translation": <float> [3] -- coordinate system origin in meters
"rotation": <float> [4] -- coordinate system orientation as quaternion
},
"gt_path": <str> -- corresponding 3D voxel gt path, *.npz
"next": <str> -- frame_token of the previous keyframe in the scene
"prev": <str> -- frame_token of the next keyframe in the scene
}
]
}
}
}
We provide a baseline model based on BEVFormer.
Please refer to getting_started for details.
Please submit your result on our evaluation server. The submission rule can be referred to here
We define a standardized 3D occupancy prediction result format that serves as an input to the evaluation code. Results are evaluated for each sample. The 3D occupancy prediction results for a the test evaluation set are stored in a folder. The participant needs to zip the results folder and submit it to the official evaluation server.
The folder structure of the results should be as follows:
└── results_folder
├── [frame_token].npz
└── ...
The result folder contains .npz files, where each .npz file contains the labels of the voxels for the 3D grids with the shape of [200,200,16]. Pay special attention that each set of predictions in the folder must be a .npz file and named as [frame_token]
.npz. The [frame_token]
in annotations.json
is the same as the sample_token
in nuscenes. A .npz file contains an array of uint8 values in which each value is the predicted semantic class index of the corresponding grid in the 3D space.
Below is an example of how to save the predictions for a single sample:
save_path = os.path.join(submission_prefix,'{}.npz'.format(sample_token))
np.savez_compressed(save_path,occ_pred.astype(np.uint8))
We provide example scripts based on mmdetection3d to generate the submission file, please refer to getting_started for details.
The official evaluation server only accepts a single *.zip
file; you can zip the results folder as below:
zip -r occ_submission.zip results_folder
Before using the dataset, you should register on the website and agree to the terms of use of the nuScenes. All code within this repository is under MIT License.