Cityscapes Attributes Recognition (CAR)

Cityscapes Attributes Recognition (CAR) 数据集是由宾夕法尼亚州立大学和Scale AI合作创建的，旨在为自动驾驶车辆提供多类别属性数据。该数据集扩展了著名的Cityscapes数据集，为每张图像中的对象添加了额外的属性标注层。目前，数据集已标注超过32,729个实例，涵盖车辆、行人等多个类别。CAR数据集采用结构化和定制化的分类法，每个类别都有自己的一组可能属性，重点关注对开发依赖精确计算机视觉和场景理解的自动驾驶算法最有益的属性。此外，还创建了一个API，以简化CAR数据集的使用。CAR数据集的应用领域主要集中在自动驾驶技术，旨在通过详细理解场景中的对象属性，提高自动驾驶系统的决策准确性和安全性。

Cityscapes Attributes Recognition (CAR) dataset was collaboratively developed by Pennsylvania State University and Scale AI, aiming to provide multi-class attribute data for autonomous vehicles. It extends the renowned Cityscapes dataset by adding extra attribute annotation layers for objects in each image. To date, the dataset has annotated over 32,729 instances, covering multiple categories including vehicles, pedestrians and others. The CAR dataset employs a structured and customized taxonomy, where each category has its own set of possible attributes, focusing on the attributes most beneficial for developing autonomous driving algorithms that rely on precise computer vision and scene understanding. Additionally, an API has been created to simplify the utilization of the CAR dataset. The application domains of the CAR dataset are mainly centered on autonomous driving technology, with the objective of improving the decision-making accuracy and safety of autonomous driving systems via a detailed understanding of object attributes in driving scenes.