Description
This dataset comprises 837 whole-slide images (WSIs) and corresponding artifact annotations collected for the development and evaluation of BigpictureQC, a deep learning framework for whole-slide image quality control in digital pathology. The dataset contains human diagnostic material and preclinical toxicology samples from multiple centers, species, tissue types, staining protocols, and scanner models. Dense pixel-level annotations were generated within selected regions of interest (ROIs). Annotated labels include artifact-free tissue/background; quality-degrading artifacts such as tissue folds, air bubbles, dust, and out-of-focus regions; and separate classes for intentional ink and marker/pen markings. Annotations are exhaustive within the annotated ROIs but should not be interpreted as exhaustive ground truth outside those ROIs.
Dataset Overview
| Title | BigpictureQC Artifact Segmentation Dataset |
|---|---|
| Short name | BigpictureQC_artifact_segmentation |
| Anatomical site |
Multi-organ / multiple anatomical sites across human diagnostic material and preclinical toxicology specimens Epithelial, lymphoid, gastrointestinal, hepatopancreatobiliary, genitourinary, respiratory, skin, musculoskeletal, nervous, ocular, glandular, endocrine, and cardiovascular tissues |
| Animal species |
Human Dog Mouse Rat |
| Specimen types |
Biopsy specimens Surgical resection specimens Diagnostic archive tissue sections Preclinical toxicology tissue specimens |
| Image type |
Whole-slide image |
| Staining |
Hematoxylin and eosin (H&E)
Immunohistochemistry (IHC; multiple markers)
Periodic acid-Schiff (PAS) and other special histochemical stains where available
|
| Cite as | [D’Amato et al., BigpictureQC: Multi-Stain, Multi-Species Artifact Segmentation for Whole-slide Quality Control [update after publication]] |
| Keywords |
quality control artifact segmentation |
| DOI | To be assigned / not yet available |
| Number of WSIs | 837 |
| Annotations | 837 |
Terms of Use
| Access approval process | This dataset is currently available under controlled access. Until the end of the BigPicture project, researchers may request access to the whole-slide images and annotations via email by using the button “Apply for access”. Requests are reviewed and, if approved, access is granted under a data-use agreement. After the end of the BigPicture project, the dataset will become available for download through the BigPicture repository. |
|---|