CompCars 汽车图像数据集

CompCars 全称 Comprehensive Cars，这是源于网络和监控两种场景下的汽车图像数据集。网络数据包含 163 个汽车品牌和 1,716 个汽车型号，监控数据包含了 5 万张汽车的正视图。该数据集总共有 136,726 张完整汽车的图像，27,618 张汽车零件的图像。完整汽车的图像被标注了边界框和视角。每辆汽车都有 5 个属性，分别是最大速度、排量、车门数量、座位数量和汽车类型。

CompCars, whose full name is Comprehensive Cars, is a vehicle image dataset collected from two scenarios: web and surveillance. The web-derived subset includes 163 car brands and 1,716 car models, while the surveillance subset contains 50,000 frontal-view car images. In total, this dataset consists of 136,726 complete vehicle images and 27,618 car part images. All complete vehicle images are annotated with bounding boxes and viewing angles. Each vehicle is associated with five attributes, namely maximum speed, engine displacement, number of doors, number of seats, and vehicle type.

搜集汇总

背景与挑战

背景概述

CompCars汽车图像数据集是一个综合性车辆数据集，包含网络和监控两种场景下的图像，总计超过16万张完整汽车和零件图像，并标注了边界框、视角及5个车辆属性。该数据集支持车辆细粒度分类、属性预测和车型验证等任务，适用于图像排序、多任务学习和3D重建等领域，但仅限非商业用途。