This notebook explores the PubLayNet dataset. PubLayNet is a large dataset of document images from PubMed Central Open Access Subset. Each document’s layout is annotated with both bounding boxes and polygonal segmentations.