Two-phase approach for deblurring images corrupted by impulse plus gaussian noise

Citations Over TimeTop 10% of 2008 papers

Abstract

The restoration of blurred images corrupted with impulse noise is adifficult problem which has been considered in a series of recentpapers. These papers tackle the problem by using variational methodsinvolving an L1-shaped data-fidelity term. Because of this term, therelevant methods exhibit systematic errors at the corrupted pixel locations and require a cumbersome optimization stage. In this work wepropose and justify a much simpler alternative approach whichovercomes the above-mentioned systematic errors and leads to muchbetter results. Following a theoretical derivation based on asimple model, we decouple the problem into two phases. First, weidentify the outlier candidates---the pixels that are likely to becorrupted by the impulse noise, and we remove them from our data set. In asecond phase, the image is deblurred and denoised simultaneouslyusing essentially the outlier-free data. The resultant optimizationstage is much simpler in comparison with the current fullvariational methods and the outlier contamination is more accuratelycorrected. The experiments show that we obtain a 2 to 6 dB improvementin PSNR. We emphasize that our method can be adapted to deblurimages corrupted with mixed impulse plus Gaussian noise, and henceit can address a much wider class of practical problems.

Related Papers

• → Nonlocal Variational Image Deblurring Models in the Presence of Gaussian or Impulse Noise(2009)34 cited
• → Image Restoration Under Significant Additive Noise(2007)21 cited
• A Comparative Approach for Analysis of Image Restoration using Image Deblurring Techniques(2015)
• → Image Restoration Method for Non-uniform Blurred Images(2018)3 cited
• Efficient methodologies for single-image blind deconvolution and deblurring(2014)