Histogram Filters for Noise Reduction
Andreas Wrangsjö, Hans E. Knutsson · KTH Publication Database DiVA (KTH Royal Institute of Technology) · 2003
A class of filters based on histograms are presented. The signal probability density function is estimated and filtering is performed in the pdf domain. Such filters can be designed to preserve signal features such as sharp edges while suppressing stochastic variations. One particular histogram filter scheme is evaluated and compared to a median filter and a normal gaussian blurring filter.