A Method for Character String Extraction from Binary Images Using Hough Transform
Osamu Shiku, Akira Nakamura, Hideaki Takahira, Hideo Kuroda · 1996
This paper describes a method for extracting character strings which have various directions and sizes. In this method, we regard a character string as a line segment which width is equal to the character size, and extract the line segment (character string) using a robust line detection method, Hough Transform. The voting scheme of Hough transform is improved from black pixel-base to line segmentbase, which contributes to improvement of extracting accuracy and reduction of processing time. In order to estimate performance of our method, it was applied to 15 free formed images (512 x 512 pixels) involving about 60 strings. As a result, almost character strings were correctly extracted. 1