Tumor region annotation for computational pathology

It marks histology regions assigned to tumor or related tissue classes under an expert annotation protocol, creating evidence for computational pathology model training.

This workflow uses histology and whole-slide tissue images. Radiology lesion segmentation uses medical scans such as CT or MRI and a different imaging context.

Yes. Define the region classes required by your protocol and use the same ontology throughout annotation and review.

Yes. Add separate classes for those regions when they are part of the task and provide clear examples in the annotation guidelines.

Yes. Define separate region classes when your computational pathology task needs the tumor boundary and its neighboring stromal context.

Provide explicit inclusion rules, annotate uncertainty where appropriate, and have specialists resolve difficult regions during review.

Supported inputs include SVS, AVS, NDPI, SCN, BIF, SVSLIDE, tiled TIFF, and single-file OME-BigTIFF slides. Tiled navigation lets teams inspect large slides at different zoom levels.

Reviewers inspect cellular or tissue boundaries in slide context, discuss uncertain regions, and return corrections through the configured workflow. Shared classes keep the annotation protocol consistent.

Yes. Pathology labels retain the original full-resolution raster coordinate space. Supported formats include COCO, YOLO, and Unitlab Unified Export Format, depending on the annotation data you need.