Scene text detection suitable for parallelizing on multi-core

Text recognition in natural color images is not successful with the nai¿ve application of general OCR (optical character recognition) because the numerous visual environments and luminance in the real world scene. This problem can be solved by adopting various brightness/contrast adjustments and texture analysis to create a binary image very similar to printed text. But such a solution can have a shortcoming; increased execution time. We propose an English text detection method that parallelizes processing jobs to utilize a multi-core processor effectively in order to improve text recognition accuracy and minimize execution time overhead.