Title :
Dynamic threshold algorithm for removal of Back-to-Front noises of visual document image
Author :
Tian, Da-Zeng ; Wang, Chao ; Zhang, Zhi-ming
Author_Institution :
Fac. of Phys. Sci. & Technol., Hebei Univ., Baoding, China
Abstract :
As we all known, the Back-to-Front noises may affect the feature extraction and classification of text when using ORC to identify. In this paper, we take advantage of dynamic threshold method to process the vision document image of Back-to-Front noise with unimodal or bimodal histogram characteristic. Experimental results show that this algorithm has a better effective.
Keywords :
document image processing; feature extraction; image classification; image denoising; image segmentation; optical character recognition; text analysis; ORC; back-to-front noise removal; bimodal histogram characteristic; dynamic threshold algorithm; feature extraction; optical character recognition; text classification; unimodal histogram characteristic; visual document image; Algorithm design and analysis; Entropy; Heuristic algorithms; Histograms; Interference; Noise; Visualization; Back-to-Front noises; Dynamic threshold; Visual document image;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2011 International Conference on
Conference_Location :
Guilin
Print_ISBN :
978-1-4577-0305-8
DOI :
10.1109/ICMLC.2011.6017030
