Create a refinement
Fork a new, editable segmentation from an existing one, seeded with a copy
of its annotations. Use this to iterate on a model prediction without
mutating the original — the new segmentation’s parent_segmentation_id
points back to the source. Scoped to your organization — a parent
segmentation outside it returns 404.
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Path Parameters
Response
Successful Response
A segmentation's metadata: where it came from, whether its annotations are
ready, and its ROI count. annotation_processing_status == ready is the
signal the mask/annotations are usable.
The mask's storage location is intentionally omitted: it is an internal S3
path (bucket + key layout) with no client value, and results are downloaded
through the export/* endpoints. Keeping it server-side avoids leaking the
bucket name and key scheme to public API-key callers.
model_prediction, refinement creating, ready