Interested in this question: are there open source libraries, services, and etc., which could highlight, give the coordinates of words or just objects without text recognition.

something like this

    2 answers 2

    Look towards OpenCV - a powerful tool for working with graphics and not only.

      Found another simple option - use the library "Tesseract". Tesseract OCR API is able to recognize text and display its coordinates. For this:

      1. We connect the library Tess4J Tesseract For Java
      2. Download traineddata , here is Russian .

      3. A simple code for an example:

        Tesseract instance = Tesseract.getInstance (); // we put the traineddata along the path "C: / TESSERACT / tessdata" // and in the "setDatapath" we specify the incomplete path instance.setDatapath ("C: / TESSERACT /"); // specify the language instance.setLanguage ("rus"); // File page page = new File ("path); BufferedImage image = ImageIO.read (page); // get the text coordinates in the form Rectangle = {x: 12, y: 34, height: 60, width: 120} / / size parameter is responsible for the size of blocks // the larger the size, the smaller the block will beat with the text