Car Parts and Car Damages

Humans in the Loop is proud to share our newest dataset for car part and car damage analysis AI systems for automated claims processing. The dataset consists of a total of 1812 images, fully annotated with polygons either for car parts (998 images) or car damages (814). The total number of polygons in the dataset is 24,851. The images were segmented by the trainees of Beetroot Academy as part of their pilot with Humans in the Loop, targeting internally displaced people across Ukraine.

This Car part and car damage dataset is dedicated to the public domain by Humans in the Loop under CC0 1.0 license.

循环中的人类自豪地分享我们最新推出的汽车零部件及汽车损坏分析AI系统数据集，旨在自动化索赔处理。该数据集包含总计1812张图像，全部采用多边形进行标注，其中998张图像用于汽车零部件的标注，814张图像用于汽车损坏的标注。数据集中多边形的总数为24,851。图像的分割工作由Beetroot学院的学员在循环中的人类公司试点项目中完成，旨在为乌克兰境内流离失所的人提供帮助。本汽车零部件及汽车损坏数据集由循环中的人类公司无偿奉献给公共领域，并按照CC0 1.0许可协议进行授权。

搜集汇总

背景与挑战

背景概述

该数据集由Humans in the Loop发布，专注于汽车部件和损坏分析，用于支持自动化索赔处理的AI系统。它包含1812张图像，其中998张标注汽车部件，814张标注汽车损坏，总计24,851个多边形注释；数据由乌克兰境内流离失所者通过培训生标注，并基于CC0 1.0许可证公开，便于公众免费使用。