CPPE-5

该数据集是一组用于研究医疗个人防护装备次级分类的具有挑战性的数据集，其中包含在复杂场景中的非标志性图像，场景中有多个物体。该数据集包括5个物体类别（防护服、面罩、手套、口罩和护目镜），并带有边界框和正面标签的注释。为了尽量减少标志性图像，图片主要从Flickr收集。该数据集规模较大，5个物体类别下拥有大量的图像。其任务是进行医疗个人防护装备的对象检测和分类。

This is a challenging dataset for researching fine-grained classification of medical personal protective equipment (PPE), which contains non-iconic images captured in complex multi-object scenarios. It covers 5 object categories: protective clothing, face shields, gloves, surgical masks, and goggles, with annotations including bounding boxes and positive labels. To minimize iconic images, the images are mainly collected from Flickr. This dataset has a large scale, with a substantial number of images across the 5 object categories. Its core task is object detection and classification of medical personal protective equipment.