An effective method in text detection and characters recognition using fuzzy theory

In this paper an algorithm for text detection and character recognition in colored images with the large number of colors is presented. Some past schemas previously proposed in the literature were dependent on histogram shape in which they did not perform well in case of vicinity on text color and background color. Although there where approaches solving this problem but they created new limitations for image type such as text character size, not using characters with different colors or not writing text characters dispersedly within image. The approach presented in this paper not only includes the benefits of previous approaches, but also removes the limitations which they created. Besides it increases accuracy in recognition step using fuzzy logic. On the other hand by combining the detection and recognition steps, it improves the speed of algorithm execution than to past algorithms.