TEST THE EFFECT OF ENHANCEMENT FILTERS ON THE NOISY TEXT IMAGE

Abstract

This paper explains an important feature of OCR (Optical Character Recognition) software and makes a comparison between them and shows how the enhancement filters effected on noisy text image, using a Hybrid Method (HM) with verifying factor with Gaussian filter gives a good result with BNTI (Black Noisy Text Image) removal, also with hybrid mean filter with verifying factor gives noise and blurring removal.

 Also the results show that with EBNTI (Enhanced Black Noisy Text Image)   the number of error is reduced than actually exit in the original BNTI and median filter work best with BNTI image than with WNTI (Wight Noisy Text Image) but still these filter causes some noise removal but do not enhance all the defected characters