No support to UNLV files

[brnrc]

When I run tesseract via cli, it automatically uses any .uzn file with the same name of the target image.

Example:
tesseract workingimage004.png output.txt -psm 4

If I have a workingimage004.uzn file in the same location of my workingimage004.png it will use it.

Tess4J ignores the uzn file.

Here is my code:

        Tesseract instance = Tesseract.getInstance();
        //In case you don't have your own tessdata, let it also be extracted for you
        File tessDataFolder = LoadLibs.extractTessResources("tessdata");

        //Set the tessdata path
        instance.setDatapath(tessDataFolder.getAbsolutePath());

        instance.setPageSegMode(4);
        try {
            String content = instance.doOCR(page);
        } catch (TesseractException e) {
            e.printStackTrace();
        }

Additional info:
Found the function that read and segments the image using a file UNLV at ccstruct\blread.cpp:36 (on tesseract source)

[4F2E4A2E]

Please provide the exact mimetype and a test file, thank you.

[brnrc]

image: https://www.dropbox.com/s/dtb1ra7jqdk9shw/workingimage004.png
uzn: https://www.dropbox.com/s/o2j6atqnwqx7kir/workingimage004.uzn

Run with: tesseract workingimage004.png manual_ocr.txt -psm 4
the output should be

ANNEX TO THE EUROPEAN SEARCH REPORT
ON EUROPEAN PATENT APPLICATION NO.

[brnrc]

Hey @4F2E4A2E,
Just a friendly question, are you checking on this?
I just want to know if should I wait or try a different solution! Thanks

[4F2E4A2E]

Yes i am, pls get online in skype, thx.

[4F2E4A2E]

Something is not right:

• the png file provided have a lot more content then two lines
• the uzn file provided seams to be empty or with just 26Bytes
• you did not provide the mimetype

Help me out, this kind of format is new to me, in order to help i must be able to be 100% sure of the data integrity of those files and i need to be able to produce the uzn files, thanks.

[4F2E4A2E]

via brucardoso2: https://github.com/OpenGreekAndLatin/greek-dev/wiki/uzn-format

[4F2E4A2E]

Thanks brucardoso2 for introducing me to uzn.
Our lib does support it, but not automated, but it will in our next release. Untill then just parse your uzn and populate the java.awt.Rectangle. Here an example without reading in the uzn file and parsing it:

File imageFile = new File("workingimage004.png");
Tesseract instance = Tesseract.getInstance();  // JNA Interface Mapping
File tessDataFolder = LoadLibs.extractTessResources("tessdata");
instance.setDatapath(tessDataFolder.getAbsolutePath());

//parse the uzn file and create a rectangle obj
Rectangle rect = new Rectangle(600, 213, 1142, 240);
    try {
        String result = instance.doOCR(imageFile, rect);
        System.out.println(result);
    } catch (TesseractException e) {
        System.err.println(e.getMessage());
    }
}

[4F2E4A2E]

A proposal has been created as you can read here: #11
@brucardoso2: can i close this issue?

[brnrc]

@4F2E4A2E sure

[4F2E4A2E]

Closing issue.