A new suppresser for impulse noise based on adaptive growth

In this paper, we proposed a swift impulse noise suppresser based on adaptive growth algorithm (AGA). It is mainly composed of three parts: signal detector, adaptive growth and indispensable reestablishment and correction. Firstly, we designed a signal detector to detect the real signal pixels and calculated signal density according to the analysis of the characteristic of impulse noise and the distribution rule of gray values of nature images. Then, we let each real signal pixel grow into a signal area whose size is confined by a function of signal density. The final restored image will be obtained nearly after a simple normalization and correction on the after-growth image matrix. Experimental results and analysis indicate that the proposed algorithm exceeds the current ones in both noise removal and computational complexity.