Illumination Compensation Algorithm for Unevenly Lighted Document Segmentation

Ju Zhiyong (University of Shanghai for Science and Technology)
Wang Pingdian (University of Shanghai for Science and Technology)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 11, No 7: July 2013

 For the problem of segmenting the unevenly lighted document image, this paper proposes an illumination compensation segmentation algorithm which can effectively segment the unevenly lighted document. The illumination compensation method is proposed to equivalently convert unevenly lighted document image to evenly lighted document image, then segment the evenly lighted document directly. Experimental results show that the proposed method can get the accurate evenly lighted document images so that we can segment the document accurately and it is more efficient to process unevenly lighted document  images than traditional binarization methods. The algorithm effectively overcomes the difficulty in handling uneven lighting and enhances segmentation quality considerably. DOI: http://dx.doi.org/10.11591/telkomnika.v11i7.2932