Constant Time Hardware Architecture for a Gaussian Smoothing Filter

Author(s):  
Ghattas Akkad ◽  
Rafic Ayoubi ◽  
Antoine Abche
2012 ◽  
Vol 485 ◽  
pp. 56-59
Author(s):  
Liang Xiu Zhang ◽  
Yong Cao ◽  
Feng Ping Cao ◽  
Rong Chun Guo

A feasible driver fatigue monitoring system based on controllable infrared light is presented. This system adopts controllable infrared light to get the light and dark image of the pupil, and preprocess the image using differences, gaussian smoothing filter and binaryzation. Then the location of driver’s eyes is detected by template matching algorithm. Finally, Kalman Filter is implemented to track the eyes in order to detect the state of the eyes. This mechanism features accuracy, responsiveness and adaptability to various situations.


2013 ◽  
Author(s):  
Irina Vidal-Migallon ◽  
Olivier Commowick ◽  
Xavier Pennec ◽  
Julien Dauguet ◽  
Tom Vercauteren

This document describes an implementation for GPU and CPU of Young and Van Vliet’s recursive Gaussian smoothing as an external module for the Insight Toolkit ITK, version 4.* www.itk.org. In the absence of an OpenCL-capable platform, the code will run the CPU implementation as an alternative to the existing Deriche recursive Gaussian smoothing filter in ITK.


Author(s):  
Hadise Ramezani ◽  
Majid Mohammadi ◽  
Amir Sabbagh Molahoseini

The two-dimensional Gaussian smoothing filter (2D-GSF) is one of the most useful techniques in image processing. Since the 2D-GSF requires high computational resources, its efficient design and implementation are critical in real-time processing purposes. Approximate computing is a new method that can be used to increase the performance of 2D Gaussian filter design with low computing overhead on field-programmable gate arrays (FPGAs). This study aims to provide a low-latency Gaussian filter architecture on FPGA such that it can be used in real-time processing applications. In this regard, accurate and approximate carry-save adders (CSAs) have been used in adder tree-based Gaussian filters. In our proposed method, we use two approximation steps: in the first step, we use an approximation structure named Speed–Power–Area–Accuracy for Gaussian filter design and in the second stage, we use approximate CSAs to convert adder-tree structures that are used in Gaussian filter, and as a result, we have significantly reduced the delay. The results of simulation and implementation show that the latency has reduced in a 3[Formula: see text] 3 2D-GSF architecture up to 22% using proposed accurate CSAs and 45% using proposed approximate CSAs, compared to existing Gaussian filters with an adder tree structure.


Author(s):  
Ibrahim F. J. Ghalyan ◽  
Ziyad M. Abouelenin ◽  
Gnanapoongkothai Annamalai ◽  
Vikram Kapila

High Efficiency Video Coding (HEVC) adopts new techniques to reduce bit-rate by 50% over a previous video compression standard. The number of intra prediction modes in HEVC is 35 modes and increased compared with the compression. Therefore, hardware architecture with common equation and a fast filter coefficient generation algorithm is proposed for low complexity intra prediction hardware. The proposed architecture performs a smoothing filter, interpolation filter, generation of predicted pixels with only Common Operation Unit (COU). Various equations in intra prediction for smoothing filter of reference samples, calculating the average of the reference samples, generating predicted pixels and filtering predicted pixels is modified to one common equation. The common operation unit using a common equation in intra prediction hardware architecture reduces hardware area and the number of computational operators to perform various equations. COU uses 2 multipliers, 9 adders, 3 shifters and generates 1 predicted sample in planar mode and 2 predicted samples in the other mode. Also, COU generates 2 filtered reference samples in filtering operation of reference samples and the average of 4x4 PU in DC mode. The fast filter coefficient generation algorithm reduces processing time by using only Look-Up Table (LUT) and adders, instead of multiplying operation and the number of computational operators. The number of gates of the architecture is 45.6k. The number of gates in the proposed intra prediction hardware is 36.7% less than previous architecture.


Sign in / Sign up

Export Citation Format

Share Document