Occluded-COCO Vehicles

该数据集名为Occluded-COCO Vehicles，是从MS-COCO数据集中选取的真实世界图像组成，其中包含了四个不同遮挡等级的图片。该数据集共分为四个遮挡级别，测试中使用了相同的六个类别（飞机、自行车、公交车、汽车、摩托车和火车）。该数据集包含了12个不同的类别，其任务是部分遮挡图像识别。

The dataset named Occluded-COCO Vehicles is constructed from real-world images selected from the MS-COCO dataset, which includes images across four distinct occlusion levels. The dataset is divided into four occlusion levels, and the same six categories (airplane, bicycle, bus, car, motorcycle, and train) are employed for testing. This dataset covers 12 distinct categories, and its task is object recognition in partially occluded images.

搜集汇总

背景与挑战

背景概述

Occluded-COCO Vehicles数据集用于部分遮挡图像识别研究，支持BIFRNet模型的训练和测试，旨在提升对遮挡车辆图像的识别能力。数据集需从指定链接下载，具体规模和内容未详细说明。