High Density Noise Removal in Endangered Silent- Films Using Non-linear Filters: A Comparative Study

Document Type : Original Article

Authors

1 Faculty of Engineering, Alexandria University, Alexandria, Egypt.

2 Faculty of Engineering, Alexandria University, Alexandria, Egypt

3 Faculty of Engineering, Pharos University in Alexandria, PUA Alexandria, Egypt

Abstract

The loss and distortion of the feature films at films silent-era constitutes an alarming in cultural recording for most nation's. So the first directive of congress library was to support archival research projects that would restore the American movies produced during the 19th and 20th centuries. Noise removal is an important task in video restoration where the "original" video is discernible. Therefore this work contemplates to introduce an overview of some algorithms that can be used to restore the corrupted video that presented in AVI container format. These algorithms are untrimmed decision based median filter (UDBMF), decision based median filter (DBMF), weighted median filter (WMF), and standard median filter (SMF). A comparable analysis among all filters was done at different levels of noise based on some calculated parameters as mean square error (MSE), mean absolute error (MAE), image enhancement factor (IEF), peak signal to noise ratio (PSNR), and correlation ratio (CORR). Simulation results indicate that DBMF and UDBMF have superior performance compared to other nonlinear filters for noise level up to ninety percent. Also the necessary details in video were preserved.