Tess4j result iterator

Question

I have the following code:

public String getName(BufferedImage subc){
        String name=null;
        Tesseract1 instance = new Tesseract1();
        instance.setPageSegMode(8);
        instance.setLanguage("eng");
        instance.setTessVariable("tessedit_char_whitelist", "qwertyuiopasdfghjklzxcvbnmQWERTYUIOPASDFGHJKLZXCVBNM0123456789_.");
        try {           
            name=instance.doOCR(subc);
        } catch (TesseractException e) {System.err.println(e.getMessage());}
        name=new StringTokenizer(name,"\n").nextToken();
        return name;
}

where subc is the image already cut and preprocessed of the word. What I want is either to obtain the confidence of the recognition of the image or to iterate the first, lets say, 30 most likely words. I have found examples like this Tess4J: How to get a Character's confidence value? , but it breaks at the first line,

TessResultIterator ri = TessAPI1.TessBaseAPIGetIterator(api);

when I put my object "instance" as the parameter "api", and after some trying to use getpointer and different objects I ve had no luck so far. Here http://tess4j.sourceforge.net/docs/docs-1.0/net/sourceforge/tess4j/package-summary.html , in the class summary I understand that maybe the objects Tesseract or Tesseract1 are not the most appropiate for what I want to do, but I didn't manage to recognize a word from an image with TessAPI or TessAPI1. The ResultIterator in c++ looks pretty concise, but with pointers: https://code.google.com/p/tesseract-ocr/wiki/APIExample Thanks!

Answer 1

The Tesseract is a simplified API, exposing only the most commonly used methods from TessAPI interface. To get the text confidence, you'll need to work with the TessAPI . The library's unit tests include some common use cases. You definitely want to take a look at them.