Combining Fuzzy C-means Clustering and Flood Filling Algorithm for Enhancing Text Binarization
Huy Phat Le, Toan Dinh Nguyen, Jonghyun Park, GuessSang Lee · 한국멀티미디어학회 학술발표논문집 · 2009
Text binarization is an important step in text understanding due to the fact that ORC (Optical Character Recognition) system only understands the binarized image. The more accurate the binarized text is; the better result the ORC system works. In this paper, a novel binarization method is proposed to binarize text from complex color images. First, the Fuzzy C-means algorithm is used to group similar pixels color in an image. The flood filling algorithm is then used to remove noise components in the background. Finally, we used Otsu global binarization method to binarize the text image. The experimental results show that our method outperforms the Otsu global binarization method.