CCAP: A Dataset with Capillary-Level Annotation and Vessel Caliber Differentiation

CCAP (Coronary Capillary and Caliber-Aware Angiography) is a re-annotation dataset for coronary X-ray angiography, providing high-quality vessel segmentation masks with an emphasis on capillary-level detail and vessel-caliber awareness. This release contains only the derived annotations and accompanying metadata, enabling research on coronary vessel segmentation, microvasculature analysis, and caliber-aware modeling in challenging angiographic conditions (low contrast, overlap, and discontinuities). To comply with the original dataset’s redistribution policy, this Zenodo record does not include any original angiography images. Instead, we provide: (1) the re-annotated segmentation masks, and (2) a mapping file (CSV) that links each mask to its corresponding image identifier/path in the original dataset, so that users can reproduce the image–mask pairing after obtaining the original images from their official source.