Abstract:Aiming at facilitating a real-world, ever-evolving and scalable autonomous driving system, we present a large-scale dataset for standardizing the evaluation of different self-supervised and semi-supervised approaches by learning from raw data, which is the first and largest dataset to date. Existing autonomous driving systems heavily rely on `perfect' visual perception models (i.e., detection) trained using extensive annotated data to ensure safety. However, it is unrealistic to elaborately label instances of all scenarios and circumstances (i.e., night, extreme weather, cities) when deploying a robust autonomous driving system. Motivated by recent advances of self-supervised and semi-supervised learning, a promising direction is to learn a robust detection model by collaboratively exploiting large-scale unlabeled data and few labeled data. Existing datasets either provide only a small amount of data or covers limited domains with full annotation, hindering the exploration of large-scale pre-trained models. Here, we release a Large-Scale 2D Self/semi-supervised Object Detection dataset for Autonomous driving, named as SODA10M, containing 10 million unlabeled images and 20K images labeled with 6 representative object categories. To improve diversity, the images are collected within 27833 driving hours under different weather conditions, periods and location scenes of 32 different cities. We provide extensive experiments and deep analyses of existing popular self/semi-supervised approaches, and give some interesting findings in autonomous driving scope. Experiments show that SODA10M can serve as a promising pre-training dataset for different self-supervised learning methods, which gives superior performance when fine-tuning with different downstream tasks (i.e., detection, semantic/instance segmentation) in autonomous driving domain. More information can refer to https://soda-2d.github.io.

A Road Surface Image Dataset with Detailed Annotations for Driving Assistance Applications

A road surface reconstruction dataset for autonomous driving

A Comprehensive Implementation of Road Surface Classification for Vehicle Driving Assistance: Dataset, Models, and Deployment

RSRD: A Road Surface Reconstruction Dataset and Benchmark for Safe and Comfortable Autonomous Driving

BAAI-VANJEE Roadside Dataset: Towards the Connected Automated Vehicle Highway technologies in Challenging Environments of China

A system of vision sensor based deep neural networks for complex driving scene analysis in support of crash risk assessment and prevention

Ithaca365: Dataset and Driving Perception under Repeated and Challenging Weather Conditions

Proximity based automatic data annotation for autonomous driving

Paved or unpaved? A Deep Learning derived Road Surface Global Dataset from Mapillary Street-View Imagery

DAWN: Vehicle Detection in Adverse Weather Nature Dataset

An Indian Roads Dataset for Supported and Suspended Traffic Lights Detection

A synthetic digital city dataset for robustness and generalisation of depth estimation models

RSUD20K: A Dataset for Road Scene Understanding In Autonomous Driving

Quality Assessment of Image Dataset for Autonomous Driving.

A Survey on Autonomous Driving Datasets: Statistics, Annotation Quality, and a Future Outlook

RoScenes: A Large-scale Multi-view 3D Dataset for Roadside Perception

Road Friction Estimation Based on Vision for Safe Autonomous Driving

Pavementscapes: a large-scale hierarchical image dataset for asphalt pavement damage segmentation

SODA10M: A Large-Scale 2D Self/Semi-Supervised Object Detection Dataset for Autonomous Driving

Opendenselane: A Dense Lidar-Based Dataset for HD Map Construction

StreetSurfaceVis: a dataset of crowdsourced street-level imagery annotated by road surface type and quality