Content retrieval from degraded document images using binarization technique

Retrieving content from degraded document image is very big and difficult task due to high variation between document background and foreground text. To simplify the task, novel system first constructs correct contrast to the image. Correct contrast to be constructed, based on the local maximum minimum algorithm process's result. Then find text edge stroke detection which lies on high contrast pixel in dark region by canny edge detection method and Otsu's method. After that, calculate local threshold value for the document image, and comparing each pixel with threshold value level, give result as `0' or `1'. Hence it is called binarization technique. Finally, post processing is done for sharpening the text edges.