CarPatch

CarPatch是由意大利摩德纳和雷焦艾米利亚大学的法拉利工程系创建的一个合成车辆数据集，专注于评估神经辐射场技术在车辆部件上的应用。该数据集包含8个不同的合成场景，每个场景对应一个高质量的3D车辆网格，具有真实细节和挑战性光照条件。数据集不仅提供RGB图像及其相机参数，还包括深度图和车辆部件的语义分割掩码，用于评估特定车辆部件的重建质量。CarPatch旨在解决车辆检查中的3D重建问题，特别是在需要详细分析车辆外观的场景，如保险评估和租车损坏责任判定。

CarPatch is a synthetic vehicle dataset created by the Department of Ferrari Engineering at the University of Modena and Reggio Emilia, Italy, focusing on evaluating the application of Neural Radiance Fields (NeRF) technology on vehicle components. This dataset contains 8 distinct synthetic scenarios, each corresponding to a high-quality 3D vehicle mesh with realistic details and challenging lighting conditions. In addition to RGB images and their corresponding camera parameters, the dataset also provides depth maps and semantic segmentation masks for vehicle components, which are used to evaluate the reconstruction quality of specific vehicle parts. CarPatch aims to address the 3D reconstruction problem in vehicle inspection, particularly in scenarios requiring detailed analysis of vehicle appearance, such as insurance assessment and damage liability determination for car rentals.