SODA10M：用于自动驾驶的大规模二维自/半监督目标检测数据集

We introduce a new large-scale 2D dataset, named SODA10M, which contains 10M unlabeled images and 20k labeled images with 6 representative object categories. SODA10M is designed for promoting significant progress of self-supervised learning and domain adaptation in autonomous driving. It is the largest 2D autonomous driving dataset until now and will serve as a more challenging benchmark for the community. Self-supervised Learning for Next-generation Industry-level Autonomous driving refers to a variety of studies that attempt to refresh the solutions on challenging real-world perception tasks by learning from unlabeled or semi-supervised large-scale collected data to incrementally self-train powerful recognition models. Thanks to the rise of large-scale annotated data sets and the advance in computing hardware, various supervised learning methods have significantly improved the performance in many problems (e.g. 2D detection, instance segmentation and 3D Lidar Detection) in the field of self-driving. However, these supervised learning approaches are notorious "data hungry", especially in the current autonomous driving fields. The performance of self-driving perception systems highly relies on the annotation scale of labeled bounding boxes and IDs, which makes them not practical in many real-world industrial applications. The intuition is that a human driver can keep accumulating experiences from self-exploring the roads without any tutor’s guidance, instead current CV solutions are still baby-sitted with extensive annotation efforts on every new scenario. To facilitate an industry-level autonomous driving system in the future, the desired visual recognition model should be equipped with the ability of self-exploring, self-training and self-adapting across diverse new-appearing geographies, streets, cities, weather conditions, object labels, viewpoints or abnormal scenarios. To address this problem, many recent efforts in self-supervised learning, large-scale pretraining, weakly supervised learning and incremental/continual learning have been made to improve the perception systems to deviate from traditional paths of supervised learning for self-driving solutions. The aim of releasing this dataset is let the public to explore methods that utilizing both labeled data and unlabled data to achieve industry-level autonomous driving solutions. The benchmark paper has been released at Arxiv and it will be used to hold the ICCV2021 SSLAD chanllege. If you have any questions about SODA10M, please contact [e-mail] or [e-mail] for further help. The annotation file keeps consistent with COCO format and contains three keys: "images", "categories" and "annotations". Image tags (i.e., weather conditions, location scenes, periods) for all images and 2D bounding boxes for labeled parts are annotated for SODA10M. - The SODA10M dataset has been released! (2021/6/8) - The SODA10M paper has been released on Arxiv! (2021/6/21) - The challenge website has been released at CodaLab! (2021/7/1) - The challenge results and technical reports have been released on Challenge page! (2021/10/21) - The SSLAD2021 workshop record video (including challenge report) has been released on YouTube! (2021/10/21) - The evaluation server has been re-opened at CodaLab! (2021/11/9)

我们提出了一款全新的大规模二维数据集SODA10M，该数据集包含1000万张未标注图像与20000张标注图像，涵盖6类典型目标类别。SODA10M旨在推动自动驾驶领域自监督学习与领域自适应技术的重大进展。它是目前规模最大的二维自动驾驶数据集，将为学术界提供更具挑战性的基准测试平台。
面向下一代工业级自动驾驶的自监督学习，指的是通过从大规模未标注或半标注采集数据中学习，逐步自训练出强大的识别模型，以革新复杂真实世界感知任务解决方案的一系列研究。得益于大规模标注数据集的兴起与计算硬件的进步，诸多监督学习方法已在自动驾驶领域的多项任务（如二维目标检测、实例分割与三维激光雷达检测）中实现性能显著提升。然而，这类监督学习方法素来以“数据饥渴”著称，尤其在当前自动驾驶领域中尤为突出。自动驾驶感知系统的性能高度依赖标注边界框与目标ID的标注规模，这使得其在诸多实际工业应用中难以落地。
其核心直觉在于，人类驾驶员可在无需导师指导的情况下，通过自主探索道路不断积累驾驶经验，而当前计算机视觉解决方案仍需针对每一种新场景投入大量标注工作，如同被全程监护的孩童。为了未来能够实现工业级自动驾驶系统，理想的视觉识别模型应当具备在多样的新兴地理区域、街道、城市、天气条件、目标类别、视角或异常场景中自主探索、自训练与自适应的能力。
为解决这一问题，近年来诸多围绕自监督学习、大规模预训练、弱监督学习以及增量/持续学习的研究工作，旨在推动自动驾驶感知系统摆脱传统监督学习的路径依赖。本数据集的发布初衷，是推动公众探索兼顾标注数据与未标注数据的方法，以实现工业级自动驾驶解决方案。
该基准相关论文已发布于Arxiv，且将用于举办ICCV2021 SSLAD挑战赛。若您对SODA10M数据集有任何疑问，请联系[e-mail]或[e-mail]获取进一步帮助。
其标注文件与COCO格式保持一致，包含"images"、"categories"与"annotations"三个键。SODA10M的所有图像均标注了图像标签（即天气条件、拍摄场景、时段），并为标注部分提供了二维边界框。
- SODA10M数据集已于2021年6月8日正式发布！
- SODA10M相关论文已于2021年6月21日发布于Arxiv！
- 挑战赛官网已于2021年7月1日在CodaLab平台上线！
- 挑战赛结果与技术报告已于2021年10月21日发布于挑战赛页面！
- SSLAD2021 workshop记录视频（含挑战赛报告）已于2021年10月21日上传至YouTube！
- 评估服务器已于2021年11月9日在CodaLab平台重新开放！

搜集汇总

背景与挑战

背景概述

SODA10M是一个面向自动驾驶的大规模二维目标检测数据集，包含1000万张未标注图像和2万张标注图像，涵盖6个物体类别，旨在促进自监督和半监督学习在自动驾驶领域的应用。它是目前最大的二维自动驾驶数据集，采用COCO数据格式，为行业级解决方案提供基准挑战。