An efficient Switching Median Filter for Salt & Pepper Impulse Noise Reduction

An improved and efficient adaptive median filter is proposed for salt & pepper impulse noise removal from digital images. The more perfect and computationally efficient filtering scheme takes care to restore only the impulse corrupted pixels by a perfect mid-ranking image statistic from an appropriate local neighborhood while keeping the signal content of the uncorrupted pixels as such. The corrupted signals are identified by an adaptive technique and the signal loss due to misinterpretation of uncorrupted signals as corrupted is almost avoided to sustain the image details intact. Experimental results show that the proposed algorithm which detects impulses and restores the corrupted signals in a single efficient stretch is far more superior than many of the other median filters in terms of retaining the fidelity of the images corrupted by salt & pepper impulse noise even to the tune of ninety five percent. The objectiveness of the proposed filter in terms of peak signal to noise ratio and mean absolute error at all noise levels shows the significance of this filter.