Our images to be processed are much like scanned document images, with white background and black text (Chinese and English letters, see the attached file), except that they are take by smart phones rather than scanners. An image typically ONLY contains several lines of text (no tables, no charts, etc.), without complex layout. I am wondering if there are any parameter for this type of image i can use to improve recognition speed?
Is there any parameter i can use to improve recognition speed?
- 929 Views
- Last Post 25 August 2014
Natalia Karaseva posted this 05 September 2014
Sure, there are several possibilities to improve recognition speed in your scenario. Please, check the recommendations from the following posts:
1257 questions, 4142 answers.