A low power selective median filter design

Dalai, Radhamadhab (2008) A low power selective median filter design. MTech thesis.

[img]
Preview
PDF
1237Kb

Abstract

A selective median filter which consumes less power has been designed and different logics for majority bit evaluation has been applied and simulated in VHDL .It is rightly called as selective because an edge pixel detector [2] has been used to select those pixels which are to be processed through median filter. As for median value calculation; sorting of 3 x 3 window’s pixel values has been done using majority bit circuit [4].Different majority bit calculation method has been implemented and the result sorting circuit has been analyzed for power analysis. In this work a general median filter which uses binary sorting method known as Majority Voting Circuit (MVC) has been designed using VHDL and optimized using SYNOPSIS which has used 0.13μm CMOS technology .The digital design of sorting circuit saves approximately 60% of power comprising of cell leakage and dynamic power comparing to a mixed signal design of Floating gate based Majority bit median filter [4]. Before operating median filter on each pixel double derivative filter [2] has been applied to check whether it is an edge pixel or not. Overall this is a digital design of a mixed filter which preserves edges and removes noises as well.Low power techniques at logic level and algorithmic level have been embedded into this work. In our work we have also designed a small microprocessor using VHDL code. Later a memory (for the purpose of image storing) based Control Unit for single median value evaluation has been designed and simulated in XILINX. Here for sorting circuit a common logic based circuit (component) has been put forward. The power, latency or delay, area of whole design has been compared and tested with other designs.

Item Type:Thesis (MTech)
Uncontrolled Keywords:VHDL, CMOS, MVC
Subjects:Engineering and Technology > Computer and Information Science > Image Processing
Divisions: Engineering and Technology > Department of Computer Science
ID Code:4342
Deposited By:Hemanta Biswal
Deposited On:11 Jul 2012 10:02
Last Modified:11 Jul 2012 10:02
Supervisor(s):Majhi, B

Repository Staff Only: item control page