CompCars

综合汽车 (CompCars) 数据集包含来自两种场景的数据，包括来自网络自然和监视自然的图像。网络自然数据包含163汽车品牌与1,716的汽车模型。总共有136,726图像捕获整个汽车，27,618图像捕获汽车部件。全车图像标有边界框和视点。每个汽车型号都标有五个属性，包括最大速度，排量，门数，座位数和汽车类型。监视性质数据包含在前视图中捕获的50,000汽车图像。

数据集可用于以下任务:

细粒度分类
属性预测
汽车模型验证
该数据集还可以用于其他任务，例如图像排名，多任务学习和3D重建。

The Comprehensive Cars (CompCars) dataset includes data from two scenarios: natural web-captured and natural surveillance settings. The natural web-captured subset contains 163 car brands and 1,716 car models, with a total of 136,726 full-vehicle images and 27,618 car part images. Full-vehicle images are annotated with bounding boxes and viewpoints. Each car model is labeled with five attributes: maximum speed, engine displacement, number of doors, number of seats, and vehicle type. The natural surveillance subset consists of 50,000 car images captured from the front view.

This dataset supports the following downstream tasks: fine-grained classification, attribute prediction, and car model verification. It can also be employed for other research tasks such as image ranking, multi-task learning, and 3D reconstruction.

AI搜集汇总

背景与挑战

背景概述

CompCars是一个大规模的汽车图像数据集，包含来自网络和监视场景的图像，其中网络数据涵盖163个品牌和1,716个模型，总计超过16万张图像，并标注有边界框、视点和五个属性（如最大速度和排量）。该数据集主要用于细粒度汽车分类、属性预测和模型验证等计算机视觉任务，发布于2015年。