question about different recognition result on same image if read some times

[huang...@gmail.com]

Dear all:

     I find one interesting problem when use tesseract 3.02. 

     I read one simple image (attatched), parameter is "char_digit-reappear-fail.bmp result -l eng -psm 10 DigitLimit". If we open ocr engine, then continue to read 3 times, result of first and second is 0, while result of third is 3. If we restart program, the result is 0. It seems that some paramter is adaptive update according to previous recognition.

     Thanks. 

[huang...@gmail.com]

This problem isresolved now.

we can call function ClearAdaptiveClassifier();

/**

 * Call between pages or documents etc to free up memory and forget

 * adaptive data.

 */