Processing blank (white) image; GetProperties on page throws exception.

[CodeWarrior-Hawaii]

I have found this to be pretty reproducible. To reproduce:

1. Call tesseract on a blank (no text) image.
2. Call Page.GetIterator().GetProperties() on the resulting Page.
3. Receive a System.AccessViolationException stating that "Attempted to read or write protected memory. This is often an indication that other memory is corrupt."

NOTE: attached file is a TIFF! Rename it from .gif to .tif. GitHub does not handle tiff images.

[charlesw]

Confirmed, the underlying issue is that your attempting to use functions on the iterator when it's not pointing to anything since there isn't any results. Normally you'd check if it's empty first using the PageIterator's Empty method. Unfortunately this method isn't yet available in the C API, as of version 3.04, so I cannot expose it. A workaround for now would be to attempt to get the text for the desired level BEFORE checking any properties. If the level doesn't exist it'll be null.

For instance:

using(var iter = page.GetIterator()) {
    iter.Begin();

    var text = iter.GetText(PageIteratorLevel.Symbol);
    if(text != null) {
         var properties = iter.GetProperties();
         // whatever...
    }
}

Obviously not ideal but I believe the only real option for now.

You can get some more ideas from the Tesseract Wiki

[CodeWarrior-Hawaii]

That will work too. In the project that I am working, I just catch the exception, spoof the properties that I need and continue on like nothing happened. I think I will go back and implement yours though. I really don't like exceptions being generated even if I do handle them.