Research Article
Denoising Epigraphical Estampages Using Nested Run Length Count
@INPROCEEDINGS{10.1007/978-3-030-20615-4_15, author={P. Preethi and K. Praneeth Kumar and M. Sumukha and H. Mamatha}, title={Denoising Epigraphical Estampages Using Nested Run Length Count}, proceedings={Ubiquitous Communications and Network Computing. Second EAI International Conference, Bangalore, India, February 8--10, 2019, Proceedings}, proceedings_a={UBICNET}, year={2019}, month={5}, keywords={Epigraphical scripts Run length count (RLC) Denoising Peak to signal noise ratio (PSNR) Structural similarity index (SSIM)}, doi={10.1007/978-3-030-20615-4_15} }- P. Preethi
K. Praneeth Kumar
M. Sumukha
H. Mamatha
Year: 2019
Denoising Epigraphical Estampages Using Nested Run Length Count
UBICNET
Springer
DOI: 10.1007/978-3-030-20615-4_15
Abstract
Denoising in epigraphical document analysis helps in building recognition system for fast and automatic processing. However, it is challenging due to the presence of stone texture as a complex background in input samples. In this paper, a nested run length counting with varying block size of 3 3, 5 5 and 7 * 7 are applied. Computation is carried out on neighboring pixels of the point of interest and discloses whether it is part of the script on inscription or background based on the count value. If it is part of the background, point of interest is set to background value else set to white. The method is tried and tested on 100 samples of epigraphical Estampages collected from archaeological survey of India. A comparative study is derived on the output of the proposed method and on the nonlinear filters such as median and wiener. Human vision perception has evaluated that proposed method is better than median and wiener filters. The quality measures such as Peak signal to noise ratio and Structural similarity indexes are practiced on the sample output for various filters and proposed method.