多尺度细节增强与自适应γ变换的图像增强

针对现有图像增强方法未能适宜地提升图像的亮度、对比度，以及保持图像自然效果的问题，提出了多尺度细节增强与自适应γ变换的图像增强方法。该方法根据多尺度的纹理结构和边缘细节特征，用引导滤波对图像进行多尺度的Retinex分解，分解为多尺度的细节层和最后的基础层；对基础层做自适应的γ拉伸，实现图像亮度和对比度的有效增强，对细节层进行多尺度的拉普拉斯增强；将增强的基础层与增强的细节层进行多尺度的Retinex反变换，实现原图像的增强。图像增强实验结果显示，相对于当前的部分最新方法，所提方法的图像增强性能更好，图像增强后的信息熵和平均梯度分别比现有的方法提升大约1.2和2.8。     

基于多尺度加权引导滤波的红外图像增强方法

现有的红外图像存在细节模糊、边缘和纹理不清晰的问题。针对上述问题,本文提出一种基于加权引导滤波的红外图像增强方法。首先,将图像通过带转向核的多尺度加权引导滤波进行分层处理,得到多幅含有细节信息的细节层图像和基础层图像;接着,对细节层采用基于Markov-Possion的最大后验概率算法和Gamma校正算法对细节层进行增强;然后,对基础层采用限制对比度的自适应直方图均衡算法进行对比度拉伸,最后,进行线性融合得到增强后的图像。综合主、客观实验结果,得出本文方法具有良好的细节增强效果,处理后的图像边缘和纹理信息比较突出,且算法在信息熵（IE）,熵增强（EME）和平均梯度（AG）3个指标都有较优的计算结果。基本满足红外图像细节得到增强,边缘纹理清晰的需求。     

自适应双边滤波与方向梯度的红外图像增强

为克服现有的红外图像增强方法存在欠增强、过增强以及对比度不高等缺陷,提出了自适应双边滤波与方向梯度的红外图像增强方法。对双边滤波进行改进,加权系数自适应于平滑区域和细节区域,以作为Retinex的中心环绕函数,将红外图像分解为基础层和细节层;用改进的平台直方图均衡化对基础层图像进行增强;提出一种方向梯度算子,用其提取细节层图像的梯度图,进而对细节层图像进行非线性的自适应边缘增强。实验结果表明,相对于部分现有的方法,本文方法能更有效地提升红外图像的亮度和对比度,增强图像的视觉效果更佳。     

基于子带分解多尺度Retinex的红外图像自适应细节增强

为实现高动态范围红外图像压缩和高亮区与阴影区细节增强,提出一种基于子带分解多尺度Retinex自适应细节增强方法。利用子带分解多尺度Retinex获取三个独立光谱子带;利用引导滤波将各子带分为细节层和基础层;之后依据子带特性设计细节增强权值基函数,自适应实现红外图像细节增强;针对输出图像平滑区灰度不均匀特点,自适应求取Gamma曲线实现灰度映射。实验结果表明:经本文算法处理后图像阴影区与高亮区细节得到明显增强,全局视觉效果良好。客观测评结果表明:本文算法有效增强图像细节信息,并且与经典基于双边滤波的细节增强算法比较,本文算法耗时没有增加。     

基于图像增强和滚动引导滤波的红外与可见光图像融合

针对红外与可见光图像融合易发生热目标亮度损失、可见光图像细节信息丢失的问题，提出一种基于图像增强和滚动引导滤波的多尺度融合算法。首先，提出一种自适应图像增强方法，提高可见光图像的整体亮度，并保持细节处的对比度。然后，根据特征的不同将源图像分解为三层，采用基于引导滤波的显著性提取方法得到亮度层；利用滚动引导滤波良好的尺度感知和边缘保持特性，并结合高斯滤波得到基础层和细节层。最后，对亮度层采用像素值取大的融合规则，提出一种新的最小二乘优化方案对基础层进行融合，使用修正拉普拉斯能量和作为清晰度的度量对细节层进行融合。实验结果表明，与其他融合方法相比，所提方法在主观评价和客观评价上都有较好的表现。     

基于改进引导滤波分层技术的红外图像增强算法

针对高动态范围红外图像在压缩显示过程中容易出现对比度低、细节模糊,以及传统增强算法在处理连续多帧的红外图像时亮度跳变的问题,提出了一种基于改进引导滤波分层技术的红外图像增强算法。该算法使用改进引导滤波对图像进行分层,从而得到质量更高的原始细节层,并使用噪声掩膜技术对其去噪;基础层使用改进的直方图均衡算法提升对比度;增强后的基础层与细节层加权融合,再通过基于神经网络的自适应伽马校正得到亮度恒定的增强图像。实验表明,相较对比算法,该算法在平均梯度上提升了2.8左右、在对比度增强测量指标上提升了10左右,同时使帧间亮度方差下降到了0.1,说明该算法具有较好的细节增强效果和亮度稳定性。     

基于变换域VGGNet19的红外与可见光图像融合

针对红外与可见光图像融合中出现细节信息丢失及边缘模糊的问题,提出一种在变换域中通过VGGNet19网络的红外与可见光图像融合方法。首先,为了使得源图像在分解过程中提取到精度更高的基础与细节信息,将源图像利用具有保边平滑功能的多尺度引导滤波器进行分解,分解为一个基础层与多个细节层;然后,采用具有保留主要能量信息特点的拉普拉斯能量对基础层进行融合得到基础融合图;其次,为了防止融合结果丢失一些细节边缘信息,采用VGGNet19网络对细节层进行特征提取,L1正则化、上采样以及最终的加权平均策略得到融合后的细节部分;最后,通过两种融合图的相加即可得最终的融合结果。实验结果表明,本文方法更好地提取了源图像中的边缘及细节信息,在主观评价以及客观评价指标中均取得了更好的效果。     

基于形态学的可变权值匹配自适应图像增强算法

为了精确提取图像的细节特征,改善图像增强的效果,该文提出了一种基于形态学的可变权值自适应增强算法。通过构造扩展全方位多尺度结构元素,并进行top-hat变换,分解了图像不同方向不同尺度的细节。该文打破了传统形态学增强算法中各方向细节取均值融合的思想,实现了根据图像的局部灰度特性,动态调整各方向不同尺度细节的权值。在图像增强过程中,根据提取到的细节的结构化特征,构造相应的自适应增益函数,从而实现图像的自适应增强。实验结果表明,该算法能较好地利用图像的自相关性,比传统的形态学增强方法更有效地突出图像的细节信息,且能抑制噪声的放大。     

低照度短波红外图像增强算法

为了拓展非制冷短波红外探测器在弱光夜视观测方面的应用，开展了针对短波红外低照度成像的研究。提出了一种新的图像增强方法抑制图像噪声增强图像细节进而改善图像质量。使用3D降噪（3DNR（3D Noise reduction））算法，将多尺度高斯差分法结合边缘保持滤波器最大限度的分离图像高频信息与隐藏噪声，再针对图像进行自适应灰度映射。实验结果表明：该算法显著地抑制了在低照度下图像的时域噪声，丰富了短波红外图像的细节，改善了短波红外的夜视显示效果。     

基于引导滤波分层的宽动态范围红外图像细节增强算法

针对宽动态范围红外图像在视觉效果方面出现的对比度低、细节信息不凸显及整体清晰度较差问题,本文提出了一种基于引导滤波分层的宽动态范围红外图像细节增强算法。该算法采用方差决策加权引导滤波对原图作分层,得到了更接近原图的基础层和更精细的细节层。为提高基础层的对比度,首先改进CLAHE的全局剪切点提升增强效果,然后基于AC视觉显著模型指导全局和改进局部直方图的融合,合理兼顾了图像背景和目标;为有效加强细节信息,基于多尺度加权引导滤波得到了信息更全面的新细节层,接着采用梯度域导向滤波对其消噪,再由Sigmoid函数压缩强边缘并突显细微目标,最后将两层信息融合并输出。实验结果表明,该算法在主观视觉和定量指标上均强于对比算法,且自适应强,鲁棒性好。     

Mobility enhancement

This article is targeted as an introduction to the physics of strained Si and the current state of the art in uniaxial strained Si MOSFET. The first part of the article explains how strain alters the valence and conduction band of Si as well as scattering rates. This is followed by a review of state-of-the-art strained techniques being implemented in 90- and 65-nm process technologies. Finally, we conclude with a discussion of the future scalability of strained Si MOSFETs in the ballistic regime and nanoscale CMOS.     

Multiscale contrast enhancement

A novel approach for image contrast enhancement is presented. A contrast measure which accounts for spatial structure information as derived within the context of a nonlinear multiscale representation generated by anisotropic diffusion is introduced. A local contrast map is computed, which is then used to enhance the original image     

Entropic contrast enhancement

A technique for the contrast enhancement of a picture is proposed. The method is derived from the entropy concept of information theory. The originality of the algorithm rests on the use of a local contrast to define the digital entropy. The basic idea of the treatment is to enhance the contrast by transforming the global entropy. The same technique can also be used for an adaptive smoothing processing.     

Color retinal image enhancement using luminosity and quantile based contrast enhancement

Multidimensional Systems and Signal Processing - Retinal imaging is used to diagnose common eye diseases. But retinal images that suffer from image blurring, uneven illumination and low contrast...     

低频重分布与边缘增强的红外图像增强算法北大核心CSCD

针对现有的红外图像增强方法存在欠增强、过增强以及微小细节丢失等缺陷,提出了低频重分布与边缘增强的红外图像增强算法。用基于改进引导滤波的Retinex将红外图像分解为低频和高频图像。为了充分利用像素级的动态空间,对低频图像进行均匀重分布,以提升图像的亮度和清晰度;用提出的方向梯度算子对高频图像进行边缘提取,再对高频图像进行边缘增强,进一步提升图像的对比度。将经增强处理的低频和高频图像作Retinex反变换,得到增强效果的红外图像。实验结果显示,相对于部分现有方法,本文方法的增强图像的信息熵和Brenner指数更高,而NIQE指数更小,因此本文方法能更有效地提升红外图像的对比度,在增强图像纹理细节的同时更好地保持图像的自然度。     

Neural network training enhancement

In this study we explore the use of nonlinear embedding maps to expand the dimension of the input space. The efficacy of such maps to speed training and to enhance performance is illustrated through several examples. A natural connection to nonlinear synaptic interconnects is also developed.     

Peer group image enhancement

Peer group image processing identifies a "peer group" for each pixel and then replaces the pixel intensity with the average over the peer group. Two parameters provide direct control over which image features are selectively enhanced: area (number of pixels in the feature) and window diameter (window size needed to enclose the feature). A discussion is given of how these parameters determine which features in the image are smoothed or preserved. We show that the Fisher discriminant can be used to automatically adjust the peer group averaging (PGA) parameters at each point in the image. This local parameter selection allows smoothing over uniform regions while preserving features like corners and edges. This adaptive procedure extends to multilevel and color forms of PGA. Comparisons are made with a variety of standard filtering techniques and an analysis is given of computational complexity and convergence issues.     

Statistical-model-based speech enhancement systems

Since the statistics of the speech signal as well as of the noise are not explicitly available, and the most perceptually meaningful distortion measure is not known, model-based approaches have recently been extensively studied and applied to the three basic problems of speech enhancement: signal estimation from a given sample function of noisy speech, signal coding when only noisy speech is available, and recognition of noisy speech signals in man-machine communication. Research on the model-based approach is integrated and put into perspective with other more traditional approaches for speech enhancement. A unified statistical approach for the three basic problems of speech enhancement is developed, using composite source models for the signal and noise and a fairly large set of distortion measures     

Low distortion speech enhancement

An innovative approach to speech enhancement is illustrated which minimises distortion to the underlying speech during the noise-reduction process. The key to this approach lies in the identification of whether the additive noise for a particular frequency component is constructive or destructive. Once this can be identified both multiplicative and subtractive filters can be derived using the minimum mean-square error criterion. The optimal combination of the proposed multiplicative and subtractive filter is also shown     

Successive noise enhancement reduction

The capacity of direct sequence code division multiple access (DS-CDMA) systems is limited by multiple access interference (MAI). Commercial CDMA systems regarded MAI as additive noise and employed the matched filter detectors. This technique limits the number of users that can be supported in a DS-CDMA system and causes an increase in the bit error rates for active users in the system as the number of users increases or as the spreading factor decreases. Thus, advanced signal processing solutions for interference suppression are necessary. In this paper, we propose a new multiuser detector (MUD) detector technique and compare between its performance to other popular MUD detectors for the universal mobile telecommunications system (UMTS) physical layer.