An Efficient Decision based Multistage Median Filter for Reducing Impulse Noise from Blood Smear Malaria Images

Microscopic imaging method is used to improve the capability to understand and examine the microscopic world. The image generated by the microscope generally gets corrupted or distorted by impulse (salt and pepper) noise. In such cases, microscopic image processing methods are used to process, analyze, and represent the images obtained from a microscope. The impulse noise gets into the image due to sensors used in a digital microscope. This gives rise to difficulties for the classification of blood cells. In this paper, an Efficient Decision-based Multistage Median Filter (EDMMF) is suggested which is used to generate a low noise high contrast image from the acquired microscopic image using microscopic digital image processing tools. Standard and Modified filters can handle low noise levels but fail for higher ones. The EDMMF algorithm is a multistage combinational use of Morphological filter, adaptive median filter, and Improved Median filter. This filter decreases the noise maintaining the image details and the edges. For noise level in digital images less than 48%, Morphological and Efficient median filter used, whereas for higher noise levels efficient median filter used. The proposed EDMMF filter has been effectively applied to noisy microscopic malaria-infected thin blood smears. The experimental results proved that the suggested method excels in terms of Peak Signal to Noise Ratio (PSNR) and Mean Square Error (MSE) over existing versions of the Median filter.