Title :
Pipelined median filters
Author :
Breveglieri, Luca ; Piuri, Vincenzo
Author_Institution :
Dept. of Electron. & Inf., Politecnico di Milano, Italy
Abstract :
Several DSP algorithms need to remove high-frequency or impulsive noise while preserving edges, e.g., in speech and image processing applications: median filtering has been proved more effective to achieve this goal than other filtering techniques. Efficient architectural implementation for real-time applications involves a careful VLSI design, which takes into account modularity, regularity, adaptability, throughput, circuit complexity, and fault tolerance. Two new architectural approaches are presented in this paper to deal with different application and implementation constraints: the pipelined polarizing median filter and the pipelined sorting median filter
Keywords :
median filters; parallel architectures; pipeline processing; signal processing; VLSI design; adaptability; architectural implementation; circuit complexity; fault tolerance; high-frequency noise; image processing; impulsive noise; modularity; pipelined polarizing median filter; pipelined sorting median filter; real-time applications; regularity; speech; throughput; Circuit noise; Complexity theory; Digital signal processing; Filtering algorithms; Filters; Image processing; Speech enhancement; Speech processing; Throughput; Very large scale integration;
Conference_Titel :
Instrumentation and Measurement Technology Conference, 1994. IMTC/94. Conference Proceedings. 10th Anniversary. Advanced Technologies in I & M., 1994 IEEE
Conference_Location :
Hamamatsu
Print_ISBN :
0-7803-1880-3
DOI :
10.1109/IMTC.1994.352171