COOOL
收藏资源简介:
COOOL数据集是由科罗拉多大学科罗拉多斯普林斯分校的研究团队创建的,旨在解决自动驾驶系统中未标记危险检测的问题。该数据集包含200个高分辨率的行车记录仪视频,由人类标注者详细标注了各种潜在的道路危险,如野生动物、不可预测的物体和标准危险。数据集的创建过程包括使用专业的计算机视觉标注平台进行标注,并由本科生和高中生在专家的监督下完成。COOOL数据集主要应用于自动驾驶领域的危险检测和预测,旨在提高自动驾驶系统的安全性和鲁棒性。
The COOOL Dataset was created by a research team at the University of Colorado Colorado Springs, aiming to address the issue of unlabeled hazard detection in autonomous driving systems. This dataset contains 200 high-resolution dashcam videos, which have been meticulously annotated by human annotators for various potential road hazards such as wild animals, unpredictable objects, and standard traffic hazards. The dataset annotation process was carried out using professional computer vision annotation platforms, and the labeling work was completed by undergraduate and high school students under the supervision of domain experts. The COOOL Dataset is primarily applied to hazard detection and prediction in the autonomous driving field, with the goal of enhancing the safety and robustness of autonomous driving systems.

- 1COOOL: Challenge Of Out-Of-Label A Novel Benchmark for Autonomous Driving科罗拉多大学科罗拉多斯普林斯分校 · 2024年



