The Visual Object Classes (Pascal VOC) was an early benchmark dataset in computer vision. Although it has been largely supplanted by newer datasets such as COCO in academic research, its XML annotation format has gained widespread adoption across other datasets, labeling tools, and models.
