An improved trilateral filter for image denoising using an effective impulse detector
Guangyu Xu, Jieqing Tan, Jinqin Zhong · 2011
In this paper, we improve the trilateral filter for removing the mix of impulse noise and Gaussian noise by adopting a new local image statistic to identify noise pixels in image corrupted with impulse noise, new statistic is called the Rank-ordered Absolute Differences statistics based on Extremum Compression (EC-ROAD). The EC-ROAD can effectively characterize each image pixel as either an impulse or an unaffected pixel, then we incorporate it into the bilateral filter by adding a third component to the weighting function. A switching mechanism is adopted to smooth the impulse noise and Gaussian noise. To improve the performance of our method we investigate the question of selecting optimal parameters. Our method has demonstrated superior performance in suppressing impulse noise and mixed noise, and outperformed other competitive methods both in terms of quantitative measures of image restoration and qualitative judgments of image quality.