摘要
图像的清晰度主要取决于它的高频分量的能量,离散快速傅里叶变换和小波变换,是两种优秀的频域变换工具。介绍了利用离散快速傅里叶变换和小波变换分析运动图像中人脸清晰度的方法:首先利用直方图均衡化以及形态学操作对图像进行预处理;然后进行人脸检测和人脸跟踪;最后进行人脸的捕捉。对比了离散快速傅里叶变换和几种小波变换的质量和效率,实验证明基于bior基的小波变换是最有效的。
The definition of the image mostly depends on the energy of the high frequency.FFT and wavelet transform are two excellent transform tools.This paper proposes a method using FFT and wavelet transform to analyze the face definition in the motion images,which consists of three steps:Firstly,the image by Histogram equalization and morphology are processed.Secondly,face detection and tracking are processed.Finally,the clearest face are captured.Through comparing the FFT with wavelet transform,the experimental results show that the bior wavelet is the most efficient.
出处
《微型电脑应用》
2010年第4期40-41,68,共2页
Microcomputer Applications
