Highlight image filter significantly improves optical character recognition on text images - Emerging Trends in Image Processing, Computer Vision, and Pattern Recognition - page 105

CHAPTER 9
Highlight image ilter
signiicantly improves optical
character recognition on text
images
Iulia Stirb Computer and Software Engineering Department, “Politehnica” University of Timisoara, Timisoara, Romania
Abstract
Image filtering is changing the appearance of an image by altering the colors of the pixels. Increasing
the contrast as well as adding a variety of special effects to images are some of the results of applying
filters, In order to obtain a high success rate of OCR (optical character recognition) performed on text im-
ages, the main target of filters is, however, to reduce the noise around characters in the image. Thereby,
I created two new nonlinear efficient image filters (for RGB and ARGB images) and elaborated the op-
timization technique suitable for each of the filters in order to make them perform faster. The first filter,
namely, Smart Contrast, increases the contrast of the image in a way depending on the value of each
component (Red, Green, and Blue) of each pixel in the image. The second image filter, called Highlight,
produces a remarkable increase in OCR success rate on the filtered text images. As Highlight filter is car-
ried out, the implementation differs from all other known filters, while the visual effect on the images
can be described as a combination between increased contrast as said before with other two visual effect:
sharpening and highlighting the edges (of the characters). Precisely, combining these specific visual ef-
fects on the resulting image makes Highlight filter so powerful in improving OCR on text images.
Keywords
Contrast
Filter
Highlight
Image
OCR