Author : Pradnya Ulhas Patil and Mr. Rahul Paikrao

Segmentation of text from badly degraded document images is a very challenging task due to the high inter/Intra variation between the document background and the foreground text of different document images. Image processing and pattern recognition algorithms take more time for execution on a single-core processor. Graphics Processing Unit (GPU) is more popular nowadays due to its speed, programmability, low cost and more inbuilt execution cores in it. The main  goal of this research work is to make binarization faster for recognition of a large number of degraded document images on GPU. In this system, we provide a new image segmentation algorithm that each pixel in the image has its own threshold proposed. We are doing parallel work on a window of m*n size and extract object pixel of text stroke of that window. The document text is further segmented by a local threshold that is estimated based on the intensities of detected text stroke edge pixels within a local window.

Keywords:  Pixel classification, GPU, Parallelization, Binarization