help with preparing an image for Tesseract OCR

[loulou2522]

I don't know how to improve the quality of an image in order to prepare an efficient Tesseract OCR ?

[DarkDragon]

I don't know how to improve the quality of an image in order to prepare an efficient Tesseract OCR ?

Try a different size, tesseract prefers high resolution scans. Screenshots are usually too low res if that's your input.

[loulou2522]

In fact no my input comes from an PDF and after i treat that with PDFTOPNG and after i submit this image to tesseract like ;
First phasen

Code: Select all

RunProgram("cmd.exe", "/C "+Chr(34)+"Pdftopng  -f "+firstpage+ " -l "+lastpage +" -gray -r 500 "+ file+ " bil"+Chr(34) ,"",#PB_Program_Wait|#PB_Program_Hide)    

Second phase

Code: Select all

RunProgram("cmd.exe", "/C "+Chr(34)+"tesseract.exe BIL-000003.png essai -l fra --psm 12 -preserve_interword_spaces=1 --dpi 500 pdf" +Chr(34) ,"",#PB_Program_Wait|#PB_Program_Hide)

third plase

Code: Select all

 RunProgram("cmd.exe", "/C "+Chr(34)+"pdftotext -f 1 -l 1 -marginl 60 -enc UTF-8 -nopgbrk -table  essai.pdf bilanactifscan.txt" +Chr(34) ,"",#PB_Program_Wait|#PB_Program_Hide)