Obtain coordinates of objects in an image

Post an image body and a comment in English and receive up to 6 sets of coordinates of given objects. You can request point coordinates as well as rectangle coordinates. Maximum image size on one side is 500 pixels. Images larger than that will be downscaled on the worker's interface.

Image-to-Coordinates captcha example, select objects on the picture

Task object

Property	Type	Required	Purpose
type	String	Yes	ImageToCoordinatesTask Defines the type of task.
body	String	Yes	File body encoded in base64. Make sure to send it without line breaks. Do not include 'data:image/png,' or similar tags, only clean base64!
comment	String	No	Comments for the task in English characters only. Example: "Select objects in specified order" or "select all cars".
mode	String	No	Task mode, can be "points" or "rectangles". The default is "points".
websiteURL	String	No	Optional parameter to distinguish source of image captchas in spending statistics.

Request example

curl -i -H "Accept: application/json" \
     -H "Content-Type: application/json" \
     -X POST -d '{
    "clientKey":"YOUR_API_KEY_HERE",
    "task":
        {
            "type":"ImageToCoordinatesTask",
            "body":"BASE64_BODY_HERE__NO_NEWLINES__NO_EXTRA_TAGS__ONLY_CLEAN_BASE64",
            "comment":"Select all elephants",
            "mode":"rectangles"
        },
    "softId": 0
}' createTask

Task solution object

Property	Type	Purpose
coordinates	String	Array of set of coordinates. For "points" mode it is (x,y) set. For "rectangles" it is (x1,y1,x2,y2), starting from top-left to bottom-right. Coordinates start in the top-left corner.

Response example

{
    "errorId":0,
    "status":"ready",
    "solution":
    {
      "coordinates":[
        [17,48,54,83],
        [76,93,140,164]
      ]
    },
    "cost":"0.000700",
    "ip":"46.98.54.221",
    "createTime":1472205564,
    "endTime":1472205570,
    "solveCount":"0"
}