Detect objects in image

API Reference

Automatically detect removable objects in an image. Use the optional lang parameter to choose the language of returned object class names.

You can pass the returned session_id and selected detected_objects item to the Remove Objects API.

Supported lang values: en (English), es (Spanish), id (Indonesian), jp (Japanese), ms (Malay), pt (Portuguese), th (Thai), vi (Vietnamese).

Model:object-detection

Credit:1

post/v1/images/detect-objects

Authorization — This endpoint requires an API key. Pass it in the api-key header on every request. Learn more

Average processing time

Per request.

Processing time: ≈ 1.2 s

Input requirements

Images

Formats: JPEG, PNG, WEBP
Maximum resolution: 3000 × 3000 px

View all input requirements

Parameters

Name	Type	Status	Description
`input_image`	file	Required	Image file or URL to detect objects in.
`lang`	string	Optional	Language for object class names.(default: `en`)

Responses

Objects detected successfullyDetectionResponse

Field	Type	Description
`session_id`	string	Session ID for the processed image. Pass as `original_session_id` to the Remove Objects API.
`detected_objects`	object[]	Detected regions sorted by confidence (descending)
`box`	number[]	Bounding box [x1, y1, x2, y2]
`mask`	string	Base64-encoded RGBG mask (same resolution as bbox)
`accuracy`	number	Confidence score (0.0 to 1.0)
`object_type`	string	Detection category
`object_description`	string	Detection class name

curl -X POST "https://api.snapedit.app/v1/images/detect-objects" \
  -H "api-key: YOUR_API_KEY" \
  -F "input_image=@./image.jpg" \
  -F "lang=en"

Response

{
  "session_id": "sess_abc123",
  "detected_objects": [
    {
      "box": [
        120,
        50,
        300,
        400
      ],
      "mask": "base64_encoded_data...",
      "accuracy": 0.95,
      "object_type": "person",
      "object_description": "person"
    }
  ]
}

Examples (1 example)

Input

Output