![]() Therefore, make sure you de-skew the text before running OCR on it. If the text is slanted or tilted, it’s harder to convert. Horizontal text is better than tilted text - OCR engines work by analyzing the document in a horizontal manner from top to bottom. Use a denoiser to reduce image noise and increase the contrast of the text alone and you’ll get more accurate conversions. You can use an image extrapolation tool to increase the resolution or dpi so you have a better chance of getting accurate OCR results.ĭenoise the document - If the text is accompanied by other meaningless characters, it makes it harder for the OCR engine to segregate actual characters from random shapes. Must be of medium or high resolution - Poor resolution text leads to poor OCR results, so make sure the images you use have the right resolution. ![]() Documents that have been scanned from wrinkled paper or images that are hazy yield poor results. Must be legible to the human eye - If you can read the document clearly, you’ll get much better OCR results. ![]() Since OCR is not always 100% accurate under all conditions, it’s better to follow some general practices before you OCR a PDF file that has been scanned or an image file containing text:
0 Comments
Leave a Reply. |