Tesseract training issue

Question

I am trying to train certain number plate images to a specific font. I'm wondering where should i put in the unicharambigs file while training it. I had a data set trained without the unicharambigs file. but when I included the tessdata folder in my code it threw an exception during run time saying "Attempted to read or write protected memory. This is often an indication that other memory is corrupt.". I am new with the use of tesseract I have a few questions to clarify because I believe the above exception is occurring because of one of the following issues

1) is it a must to have the unicharambigs file? if so where should I put in this file before creating the final trained data?

2) what files should be taken in to account in the tessdata folder when integrating with the C# code?

I have been testing with a pretrained data set but I want to have my own data set and when I use that tessdata folder the application wouldn't exit with the above mentioned run time exception. therefore I believe its something to do with my own tessdata folder.

UPDATE

Just to add to the question. When I used the default tessdata folder created by tesseract installation the application works fine but with a faulty result. I trained tesseract again and replaced the existing traineddata file with the new traineddata file that was generated after training and my application crashed again.

I am using C# and tesseract dot net wrapper.

Answer 1

Actually here problem is version mismatch of trained tessedata and tesseract DLL .I guess that trained tessedata is latest >3.0 version because of that old version (2.0) dll is throwing the exception like Attempted to read or write protected memory" so please try with latest tesseract dll . https://github.com/charlesw/tesseract/tree/master/Samples download the sample project from above link and test your trained data,Its have the reference of Nuget tesseract package so this will support Tesseract 3.02 trained tessedata.