Document Type
Thesis
Date of Award
9-30-1990
Degree Name
Master of Science in Electrical Engineering - (M.S.)
Department
Electrical and Computer Engineering
First Advisor
Chung H. Lu
Second Advisor
Yeheskel Bar-Ness
Third Advisor
Nirwan Ansari
Abstract
In the early development of signal and image processing, linear filters were the primary tools. Their mathematical simplicity made them easy to design and implement. They also offered satisfactory performance in many applications. However, they show poor performance in other cases, such as image processing.
Nonlinear filtering offered a solution to these problems. Nonlinear filtering has undergone a lot development for a long time. Among the first known nonlinear filters was the median filter, which was proposed by Tukey. Median filter preserves signal or image details better compared to a linear filter. Later other nonlinear filters were developed, including median based filters, rank-order based filter family etc.
A new filter family the Max/Min filter is proposed here. This was studied in this thesis and was found to offer better filtering compared to the classical median filter together with a reduction in processing time. The filter family has interesting operational properties, these were also investigated. Some of these properties are unique to the family while others are typical for rank order based filters. The statistical and the spectral properties for the filter output were studied with white noise input. A number of modifications on the basic filter are also suggested.
Recommended Citation
Kamel, Raafat E., "Max/Min : a novel rank order based filter family" (1990). Theses. 2774.
https://digitalcommons.njit.edu/theses/2774
