API
Image Recognition API

Recognize

Create a request for processing a group of photos.

Images are processed in batches - scenes. A report is generated for all images in the scene. In one scene, there can be photos of different parts of one shelf or photos of several shelves of the same product category.

Images must be pre-loaded into the system via /uploads or /uploads_by_url.

Note

The number of requests to the point is limited to 100 requests per minute.

POST /recognize
POST
Request
Body Parameters
images 
optional
Object
List of strings. List of image IDs for processing.
display
optional
String
List of displays ID. The field is optional, but some report types are not available without the display, e.g. Planogram Compliance and Must Have List.
report_types
optional
Object
List of strings. List of reports to be generated.
visit
optional
String
Use this field to include images to a previously created visit. The field is optional. If not specified, a new visit will be created.
datetime
optional
String
The field is optional. If you need to specify the date and time of the visit.
webhook
optional
String
Use webhook to reduce latency. If specified, the reports will be sent by POST request to this URL immediately when ready (one by one). The field is optional. If not specified, the client should explicitly poll reports.
panorama_stack_height
optional
Number
Hint to panorama stitching engine, see PANORAMA report.
panorama_stack_width
optional
Number
Hint to panorama stitching engine, see PANORAMA report.
counry_code
optional
String
Two-character сountry code to specify from which country the sku on the photo eg «RU», «KZ». Standart ISO 3166 (A2).
retail_chain
optional
String
Use retail_chain to define retail chain that owns the shop.
metadata
optional
Object
Dictionary of additional parameters.
metadata/agent_id
optional
String
Agent Client ID . If the agent exists and the name has been changed, it will be updated. The created scene will be associated with this agent.
metadata/agent_name
optional
String
Agent name. Used to create or update an agent.


In response, the system will return report identifiers. The reports can be accessed later using /reports.

If a webhook is specified, once the reports are generated, JSON will be sent to it with a POST request.

201


Response parameters



id string (ID)

Report ID.



images list of strings

List of image ID's from the request.



display integer

List of display IDs.



reports dictionary report type -> report ID.

Report ID's for the scene.



scene string

Uuid created scene ID.





Updated 03 Mar 2023
Did this page help you?