多尺度Gabor小波变换在图像检索中的应用

在基于内容的图像检索中,纹理特征是一种重要而又难以描述的特征。为提高图像检索中纹理特征的提取效率,通过对Gabor滤波器滤波特点的研究,提出一种基于多尺度Gabor小波纹理的图像检索方法。设计了一组具有多种尺度和多个方向的滤波器组,选择并优化滤波器组的各参数,对图像进行滤波和特征提取。设计并实现了一个基于Gabor纹理...     

基于Gabor滤波器医学图像的纹理分割

基于Gabor滤波器纹理分割的研究。Gabor滤波器的改进,对Gabor滤波器在方向上做了改进,把Gabor滤波器在4个方向上提取纹理特征,拓展成在6个方向上提取纹理特征,基本满足了多通道多分辨率的要求。     

基于Gabor滤波器人体血液图像中粘连细胞的分割

提出一种基于二维Gabor过滤器对人体血液细胞图像中粘连的红细胞和白细胞进行分割的方法.通过分析、简化、分解Gabor函数,设计了Gabor的多个参数,包括中心频率(ρ),滤波方向θ,以及沿着z轴和y轴Gaussian包络的空间常量σ.在算法的过程中,使用一组包括多个方向和多个频率组合的Gabor过滤器对血液图像进行滤波,提取出图像的纹理特征,然后按照一定的规则对特征图像进行合并,最后再用聚类算法进行分割.实验结果表明该方法具有很好的效果.     

结合Gabor滤波器和ICA技术的纹理分类方法

提取有效的特征用于纹理描述和分类一直是纹理分析的难点.本文提出一种结合Gabor滤波器和ICA技术的纹理特征提取方法,即纹理图像首先经过Gabor滤波器组滤波,然后由滤波图像直接构建高维特征矢量;再将这些高维特征矢量通过主成分分析PCA进行降维,最后采用ICA技术分析和提取降维后的特征矢量中的独立成分用于纹理分类.通过与经典Gabor滤波器和ICA方法的对比实验,验证和评价了本文方法的性能.     

多方向多尺度Gabor特征表示及其匹配算法

Gabor滤波是众所周知的一类特征提取方法,在机器视觉等领域得到了广泛研究和应用.本文提出了一种多方向多尺度Gabor特征表示、提取以及其匹配算法.多方向多尺度Gabor特征通过使用一组不同尺度和不同方向的Gabor滤波器对图像进行滤波,而后将滤波结果在各个滤波方向按尺度大小排序后连接而成.本文进一步提出了循环向量的概念,并将两个多方向多尺度Gabor特征相似度重新定义为一个多方向多尺度Gabor特征和对应的多个循环向量之间最大值.实验结果表明,本文提出的多方向多尺度Gabor特征不仅具有平移不变性、旋转不变性、尺度不变性,也展现出优秀的局部特征表示能力以及显著的鉴别力.     

基于Gabor变换的中文字符特征提取方法研究

针对传统汉字字符特征提取方法的不足,提出了一种基于Gabor变换,对图像纹理特征的方向性敏感的字符特征提取方法。先将灰度字体图像进行二值化、归一化处理,再利用Gabor滤波器进行滤波,获取图像的内部纹理特征和边缘信息,进而从不同方向进行特征提取。实验证明利用gabor变换计算中文字符图像的笔画方向特征,对中文字符具有较好的分类效果。     

一种基于可变长起主导作用特征(VLDF)的人脸识别算法

为了从Gabor滤波后的图像中提取简单有效、区分力强的人脸特征,提出了一种基于可变长起主导作用特征(VLDF)的人脸识别算法.即首先人脸图像与不同尺度、不同方向的Gabor滤波器进行卷积运算,然后利用局部二元模式(LBP)算子提取滤波输出的纹理特征,并根据纹理特征的统计分布规律,采用数量可变的起主导作用的纹理模式作为人脸的VLDF特征.最后构造了VLDF人脸特征之间距离的计算方法.该算法具有较小的特征向量维数和高的rank-1识别率.在FERET人脸数据库上的仿真结果验证了算法的高效性.     

基于Gabor变换的图像纹理识别

提出了一种改进的Gabor变换的图像纹理增强算法,将空域的纹理图像变换到联合空间频率域并将联合空间频率域的能量分布作为掌纹的特征,Gobar滤波器在频域内对子块其主方向上频率能量分布进行滤波,特征向量进行了匹配识别,增强特征纹线信息并对核心区域进行多方向空间频率能量滤波合成,同时也对滤波算法进行了优化,有效地减少了运算量.实验证明,该套增强算法运行稳定,效果好,可有效提取纹理图像特征纹线信息.     

一种改进的基于Gabor滤波器的纹理分割方法

本文深入分析了Gabor滤波器的性能及其实现纹理分割的原理，针对现有的基于Gabor滤波器纹理分割方法存在的不足，提出一种根据“频谱特征值向量”构造Gabor滤波器组，以实现纹理分割的改进方法。实验结果表明，本文所提改进方法的分割效果明显优于现有方法，充分证实该方法行之有效。     

分割双纹理图像的最佳Gabor滤波器设计方法

本文提出了一种用单Gabor滤波器分割双纹理图像时最佳Gabor滤波器的设计方法.该方法根据纹理图像能量在频域上的分布相对集中但并不局限于单一频率的特性,通过比较两类纹理之间的Fourier功率谱密度来设计最佳Gabor滤波器.与现有的方法相比,它具有计算量小、分割效果好的特点.另外它还提供了一种根据Gabor滤波器的频率带宽参数来指导搜索最佳滤波器的新策略,这与人类视觉系统的特性是相吻合的,具有很强的可操作性.     

Fast IIR isotropic 2-D complex Gabor filters with boundary initialization.

Gabor filters are widely applied in image analysis and computer vision applications. This paper describes a fast algorithm for isotropic complex Gabor filtering that outperforms existing implementations. The main computational improvement arises from the decomposition of Gabor filtering into more efficient Gaussian filtering and sinusoidal modulations. Appropriate filter initial conditions are derived to avoid boundary transients, without requiring explicit image border extension. Our proposal reduces up to 39% the number of required operations with respect to state-of-the-art approaches. A full C++ implementation of the method is publicly available.     

用Gabo:展开进行地震数据的滤波

Gabor展开是用一组在时域及频域都局部化且具有能量集中性质的函数来展开信号,这种特征使得Gabor展开适于处理那些时间无关或非平稳的信号。利用框架理论,类似于SVD特征映像滤波方法,本文用Gabor展开滤波方法来进行地震信号的去噪处理。仿真结果显示出Gabor展开滤波方法的优越性。     

SAR ocean image decomposition using the Gabor expansion

Demonstrates the utility of the Gabor expansion as a new tool in geophysical research. The Gabor expansion provides good time-frequency (or space-wavenumber) localization and is ideally suited to represent nonstationary processes. The properties of this tool are demonstrated by expanding an FM-chirp waveform, and azimuth cuts taken from two different SAR ocean images. The effects of filtering in Gabor phase space are also investigated     

Gabor expansion for adaptive echo cancellation

A good echo cancellation algorithm should have a fast convergence rate, small steady-state residual echo, and less implementation cost. The normalized least mean square (NLMS) adaptive filtering algorithm may not achieve this goal. We show that using the Gabor expansion is a way to achieve this goal. For direct digital signal processing compatibility the Gabor expansion introduced in this paper is for discrete-time signals, although the Gabor expansion also can be used for continuous-time signals. The Gabor expansion can be defined as a discrete-time signal representation in the joint time-frequency domain of a weighted sum of the collection of functions (known as the synthesis functions). There are several design issues in the echo canceller based on the Gabor expansion: the design of the analysis functions for the far-end speech, the design of the analysis functions for the near-end signal containing the echo plus the near-end speech, the design of the adaptive filters in the subsignal path, and the design of the synthesis functions. All the adaptive filters are designed using identical NLMS adaptive filtering algorithms     

Seismic data filtering using a Gabor representation

Gabor representations are signal expansions using sets of functions that are localized and concentrated in time and frequency. This characteristic makes them suitable candidates for filtering data where the desired signal or noise is nonstationary or time-dependent. It is shown that Gabor representations formulated with frame theory can be used for time-dependent noise removal. Furthermore, their ability to filter noise in the presence of a nonstationary signal enables them to outperform singular value decomposition eigenimage filtering techniques in the removal of incoherent noise in common midpoint data with a moderate fold     

基于Bayes准则自适应Gabor系数缩减的信号恢复

信号在联合时频平面内是相对稀疏的,鉴于此,本文通过假设信号Gabor展开系数先验分布和噪声Gabor展开系数的先验分布,然后应用Bayes准则从含噪信号Gabor展开系数的后验分布估计未知信号的Gabor展开系数,实现时变滤波.数值实验表明该方法是合理有效的.     

An unsupervised texture segmentation algorithm with feature spacereduction and knowledge feedback

This paper presents an unsupervised texture segmentation algorithm based on feature extraction using multichannel Gabor filtering. It is shown that feature contrast, a criterion derived for Gabor filter parameter selection, is well suited for feature coordinate weighting in order to reduce the feature space dimension. The central idea of the proposed segmentation algorithm is to decompose the actual segmented image into disjunct areas called scrap images and use them after lowpass filtering as additional features for repeated k-means clustering and minimum distance classification. This yields a classification of texture regions with an improved degree of homogeneity while preserving precise texture boundaries.