Skip to main content

Extract picture and images

Feature Request

Hi,

We would like a feature that allows us to retrieve a picture from a resume that has been sent to analyze via API.


The best option would be to have, in the list of extracted data from the “Resume OCR”, a new element :
{ "image": [ {“type”: “picture|logo|other”, "id": "document_unique_id", '“dimension”: “widthXheight”, “location”:”coordinates” } ] }

  • type : would describe the type of images ; eventually this could also be added as a filter in the OCR request

  • document_unique_id : would be the id of the extracted document, available for download via Mindee API

  • widthXheight : the dimensions of the picture, width by height, for example “150×450”

  • coordinates : the abscisse and ordinates of the pictures in the original documented, for example “ 300;100”

The coordinates would be usable by ImageMagick.

Thanks

Posted by Mansuy DEJEAN ·
Status: Completed8 comments

Log in to comment and vote

Comments8

  • Maxime Churin

    Team•

    Aug 14, 2025

    Hello, we just released a new field name object detection to handle such use cases. Let us know your feedback.

  • Ianaré Sévi

    Team•

    Jul 21, 2025

    Hi, this is currently not possible however we are working on adding object detection as a field type to the Data Schema. I’ll merge the requests so you get updated when the feature is live. Thanks!

  • Jonathan Grandperrin

    Team•

    Jul 1, 2025

    Hi! Thanks for your request. This is closely related to something we already have in our roadmap: https://feedback.mindee.com/en/p/object-detection-request

    We plan to give you the ability to add fields that detects and extract the location of objects in images and pdfs. For example, you’d be able to add the extraction of a “Resume owner photo” field in your data schema, that will extract the coordinates of the photo of the Resume’s owner in your example. You could also add another field “logos” if you want to extract the logos in the resume.

    The output would be the relative coordinates of the detected entity in the document, and the page number (in case it’s a multi page pdf). We don’t plan on ‘cutting’ the image and proposing a download link, but the coordinates should be enough for you to do that in your code.

    An example output for your case would be:

    {
    "resume_owner_photo": { 
        "location": [ 
            { 
                "page": 0, 
                "polygon": [0.1221, ..., 0.1223] 
            }
        ]
    }
}

    widthXheight would be easily computed using the coordinates in the polygon.

    Would that work for you?

    • Mansuy DEJEAN

      •

      Jul 3, 2025

      It looks like a good solution.

      May I ask to what correspond the coordinates in the “polygon” field ? Can they be used directly with imagemagick ?

      Thanks

      • Jonathan Grandperrin

        Team•

        Jul 3, 2025

        I don’t know well enough imagemagick but the polygon output is quite standard and should be sufficient to be used in the library to crop the image around the detected area. 

        ..."polygon": [
    [0.681, 0.481], 
    [0.779, 0.481], 
    [0.779, 0.506], 
    [0.681, 0.506] 
]...


        A polygon is a list of tuples, each tuple is a point in the image (the x and y coordinates), in relative format. Meaning for instance that [0.681, 0.481] corresponds to a point that is at 68.1% of the image on the x axis, and 48.1% on the y axis.

        From these polygons, you can pretty much reconstruct any required format for any image manipulation library, sometimes with a bit of post processing in your code, sometimes directly as-is.

        • Mansuy DEJEAN

          •

          Jul 3, 2025

          It should work fine then, thanks

  • charles

    •

    Jun 26, 2025

    Thank you Jeremy, we are working on it !

    • Jeremy

      •

      Jun 27, 2025

      Thank you for taking prompt action Charles ! Appreciate it :)