Skip to content

Object Recognition Activity

The Object Recognition activity detects and extracts multiple objects within an image using object localization. It identifies each object and provides a LocalizedObjectAnnotation containing information about the object's position and its rectangular bounds within the image.

Field Description Requirement
Api key Google Cloud Vision API Key. Required
Image path/url Google Cloud Storage URL of the image. Required
Base64 Base64 encoded string of the image file. Required
Response variable name The name of the variable to store the HTTP response. Required

Action Types & Examples

Object Recognition Result

  • Format: JSON Object (representing a collection of LocalizedObjectAnnotations)
  • Example Result: {"localizedObjectAnnotations": [{"mid": "/m/01g3x", "name": "Cat", "score": 0.95, "boundingPoly": {"vertices": [{"x": 100, "y": 50}, {"x": 200, "y": 50}, {"x": 200, "y": 150}, {"x": 100, "y": 150}]}}]}

Implementation Examples

Field Setup - Response variable name: getRespone

Execution Parameters - Api key: bb3c53c6-8515-4e34-97c8-8281de60972d - Image path/url: gs://cloud-samples-data/vision/ocr/sampleimage.jpg - Base64: /9j/4AAQSkZJRgABAQEAYABgAAD/2wBDAAMCAgMCAgMDAwMEAwMEBQgFBQQEBQoHBwYIDAoMDAsKCwsNDhIQDQ4RDgsLEBYQERMUFRUVDA8XGBYUGBIUFRT/2wBDAQMEBAUEBQkFBQkUDQsNFBQUF.../58x/8BX+R/9k=

Technical Notes

For comprehensive details on Google Cloud Vision's object localization capabilities, refer to the official Google Cloud Vision documentation. The activity requires either the Image path/url or Base64 parameter to be configured, but not both simultaneously.