A New Class of Midpoint type Nonlinear Filters for Eliminating Short and Long Tailed Noise in Images

A new class of midpoint type filters for image processing is proposed in this short paper. First, we describe Alpha Trimmed Midpoint (ATMP) filter which is quite effective in eliminating impulsive (long tailed) and uniform (short tailed) noise. Next, we introduce Adaptive Alpha Trimmed Midpoint (AATMP) filter which performs midpoint filtering in the homegeneous regions and operates as a median smoother close to the edges. Finally, a modified filter, which we call Modified Adaptive Alpha Trimmed Midpoint (MAATMP) filter, is proposed. MAATMP filter provides good noise reduction for the data confounded by uniform and impulsive noise, retains edge structures and preserves fine details. The proposed filters are evaluated for their performance on a test image and the results obtained are included.