Periphery (Camera)

Capture Semantic Segmentation

Retrieves semantic segmentation data for specified object classes, with each detected entity uniquely labeled. If no class is specified, data for all labeled objects in the scene is returned.

Use the set_semantic_label endpoint to assign labels to scene entities.

get/periphery/cameras/capture/semantic-segmentation

Query parameters

result_type'json' | 'rgb_png' | 'jpeg' required

Format which will be used to represent the captured data. 'json' carries the measurements; 'rgb_png' and 'jpeg' are renderings to look at, and jpeg is lossy.

object_classstring[]

The class of the object for which the data is to be fetched. If no labels are set in the scene, all bounding boxes could be returned.

camera_prim_pathstring nullable

Path of camera prim. Leave it out to capture through the active viewport camera.

widthinteger

Image width in px

heightinteger

Image height in px

Response

Successfully fetched semantic segmentation data

Changes

Changed in 1 of the 13 revisions of this API.5

Sep 24, 2026f4e4cbd85f5b5See the full diff
- ○
  the query request parameter camera_prim_path became optional
  request-parameter-became-optional
- ○
  added the new enum value jpeg to the query request parameter result_type
  request-parameter-enum-value-added
- ○
  query request parameter camera_prim_path list-of-types was widened by adding types null
  request-parameter-list-of-types-widened
- ○
  for the query request parameter camera_prim_path, the type was generalized from string to no type
  request-parameter-type-generalized
- ○
  added the media type image/jpeg for the response with the status
  response-media-type-added

Full history