LAVA交通标志数据集

LAVA交通标志数据集是由加州大学圣地亚哥分校创建的，包含14,112个样本，主要用于自主车辆路径规划中的显著性敏感标志分类研究。数据集包括交通标志边界框、细粒度交通标志类别标签以及二进制标签，指示交通标志的显著性。此外，还提供了每个场景的辅助元数据，包括道路类型和自主车辆的下一个计划操纵。数据集的创建旨在帮助研究人员更好地理解和预测交通标志对自主车辆路径规划的重要性，从而提高道路安全。

The LAVA Traffic Sign Dataset was created by the University of California, San Diego, and contains 14,112 samples, which are primarily used for saliency-aware sign classification research in autonomous vehicle path planning. The dataset includes traffic sign bounding boxes, fine-grained traffic sign category labels, and binary labels indicating the saliency of traffic signs. Additionally, auxiliary metadata for each scenario is provided, including road type and the next planned maneuver of the autonomous vehicle. The dataset was developed to help researchers better understand and predict the importance of traffic signs to autonomous vehicle path planning, thereby improving road safety.