改进增量式非负矩阵分解算法及其在人脸识别中的应用

非负矩阵分解(NMF)是一种有效的子空间降维方法,凭借其可解释性在人脸识别方面有着较好的应用。而增量式非负矩阵分解(INMF)利用近似的原则将上一步迭代寻优的运算结果参与后续计算,有效改善了NMF算法运算规模随训练样本增多而不断增大的现象。文章提出的改进增量式非负矩阵分解算法(Improved Incremental Non-negative Matrix Factorization)在INMF的基础上进一步利用了新加入样本的类别信息,优化了算法中参与迭代的增量系数向量的初始化值,使目标函数在迭代求解时具有更快的收敛速度和全局寻优能力。通过在ORL和YALE人脸数据库上的实验表明,该算法在运算速度和识别率上均优于传统的NMF算法和INMF算法。     

基于多原子快速匹配追踪的图像编码算法

该文提出一种多原子快速匹配追踪信号稀疏分解算法,并将其应用于静态图像编码。多原子匹配追踪通过每次迭代选取多个原子的形式,实现信号的快速稀疏分解。在此基础上,通过构造多尺度脊波字典实现图像的稀疏分解,并对稀疏分解的数据进行自适应量化和编码。实验结果表明,多原子匹配追踪获得了与匹配追踪相当的逼近性能,同时极大地提高了稀疏分解的速度。新的编码算法在低比特率情况下,获得了比JPEG2000更理想的编码性能。     

利用FFT实现基于MP的信号稀疏分解

该文研究基于Matching Pursuit（MP）方法实现的信号稀疏分解算法，通过对信号稀疏分解中使用的过完备原子库结构特性的分析，提出了一种新的信号稀疏分解算法。该算法首先通过利用原子库的结构特性，很好地处理了稀疏分解过程中计算量和存储量之间的关系。在此基础上，把信号稀疏分解中计算量很大的内积运算转换成互相关运算，最后用FFT实现互相关运算，从而大大提高了信号稀疏分解的速度。算法的有效性为实验结果所证实。     

非相干子字典多原子快速匹配追踪算法

从冗余字典中得到信号的最稀疏表示是一个NP难问题,即使是次优的匹配追踪仍然相当复杂.该文提出一种多原子快速匹配追踪算法.该算法首先将冗余字典分解成M个非相干的子字典,每次迭代分别从各子字典中至多选取一个满足条件的原子组成多原子集;最后通过求信号在多原子集上的正交投影,得到信号的多原子稀疏逼近.实验采用真实音频信号进行仿真;结果表明新的算法获得与匹配追踪相当的稀疏逼近性能,同时大大提高了信号稀疏分解的速度.     

快速TFAD在雷达辐射源信号中的应用

计算复杂度高是制约时频原子分解算法在信号处理中应用的主要问题,文中提出了一种基于粒子群算法(PSO)的时频原子分解快速算法,该方法在过完备Chirp原子库的基础上,采用时频原子分解算法分解信号,并通过PSO算法降低时频原子分解算法搜索过程的计算复杂度,提高信号处理效率.对雷达辐射源信号的仿真实验结果表明,该方法与传统的时频原子分解算法相比计算速度大幅提高,且用在数量上比Gabor少的Chirp原子刻画出信号的主要时频特征.     

迭代预测正交匹配追踪算法

正交匹配追踪系列算法中,每次迭代在原子库中选择和残差匹配的多个原子是主流的改进方向,但对多原子的选择标准却鲜有深入研究,一般是选择原子库中与残差相关系数中最大的K个原子,或者选择所有大于某一阈值的原子。本文以正交匹配追踪算法为原型,运用统计学方法,研究了相邻两次迭代中与残差相关系数最大的原子之间的关系,得出了其相关系数具有区间性的结论,这对一次迭代选择多个原子具有指导意义。该结论可以支撑对下一步迭代中的原子进行高概率预测。基于此,本文提出了迭代预测正交匹配追踪算法,实验结果表明,相对于其他匹配追踪算法,其在保证重构精度未降低的情况下,耗时有较大幅度降低。      

一种80位扩展双精度浮点三角函数运算单元的设计

三角函数求值这一运算计算过程复杂,硬件较难实现.针对这一问题,通过改进CORDIC算法,实现了兼容SPARC处理器和INTEL处理器浮点标准的80位高精度浮点三角函数的计算.在算法设计中,将函数的计算范围扩展至-π~+π,并且实现了迭代次数的可配置.最后验证了算法的正确性与完整性,分析了运算过程中迭代次数与精度的关系.结果表明,运算的精度提高到10-14.     

多模式CORDIC算法结构改进与实现

本文对计算反正余弦函数的CORDIC算法的迭代结构进行了改进,并在此基础上完成多模式CORDIC算法的实现.通过重新设定初始旋转向量避免了前两级迭代,通过修改向量旋转方向的判决条件对原算法的误差进行了校正,在增加了很少资源的情况下将正余弦运算和反正余弦运算统一到同样的迭代结构中并予以实现.实现结果表明改进后的算法反正余弦运算结果有更高的运算精度,在两种运算函数都需要的应用中能够有效减少的硬件资源占用.     

快速最小模算法新实现及实验研究

KumaresantTufts提出的最小模算法（KT）通过对空间协方差矩阵进行特征分解后构造噪声子空间向量来求解方位。VictorT．Ermolaev和AlexB.Gershman在此基础上利用指数基替代特征向量基建立了一种迭代算法（EG），避免特征分解过程，从而减少了原算法的运算量。本文对该算法再进一步改进（MEG算法），合理简化参数设置，有效解决EG算法中的参量选取和迭代收敛问题，使运算量又得到大幅度降低，更接近工程应用。文中介绍了MEG算法各参量的详细设置过程、利用计算机仿真与KT算法的性能进行统计分析比较。我们开展水池实验研究和VISI实时运算实验来验证其实用性。各种结果表明MEG算法性能优越、运算小，具有良好的应用前景。     

快速收敛的粒子群改进算法

本文提出了一种具有快速收敛速度改进算法(IPSO)。改进算法把梯度法的思想融入粒子群算法,针对算法的早熟现象,采用速度监控策略;针对算法后期粒子在最优值附近徘徊不易收敛的现象,定义了双重收敛精度。对三个BenchMark函数的测试结果表明,新算法提高了运算效率,有效的避免早熟现象的产生,并在迭代后期更有效更精确地找到测试函数极值点。     

基于多曝光的高动态图像合成的噪声处理

针对在高动态范围图像合成的过程中有噪声影响图像的质量这一问题,采取一种基于多曝光图像的高动态范围图像合成降噪算法。通过对各曝光图像的灰度数据进行提取、整理、分析,能合成代表原始场景光线分布的亮度图像。通过分析噪声对高动态范围图像合成质量的影响,提出在图像合成前将图像中含有的噪声进行处理。根据光子散粒噪声变化的特点,将图像混有的噪声问题转化为求解一个多曝光图像序列组的平均值问题,合成的图像视觉效果与真实图像极为接近。     

High-quality non-blind image deconvolution with adaptive regularization

Non-blind image deconvolution is a process that obtains a sharp latent image from a blurred image when a point spread function (PSF) is known. However, ringing and noise amplification are inevitable artifacts in image deconvolution since perfect PSF estimation is impossible. The conventional regularization to reduce these artifacts cannot preserve image details in the deconvolved image when PSF estimation error is large, so strong regularization is needed. We propose a non-blind image deconvolution method which preserves image details, while suppressing ringing and noise artifacts by controlling regularization strength according to local characteristics of the image. In addition, the proposed method is performed fast with fast Fourier transforms so that it can be a practical solution to image deblurring problems. From experimental results, we have verified that the proposed method restored the sharp latent image with significantly reduced artifacts and it was performed fast compared to other non-blind image deconvolution methods.     

基于色调处理技术的图像认证算法

基于色调处理技术,该文给出了一种有效可行且具有自修复能力的图像认证算法。首先,基于误差扩散色调处理技术将水印图像4bit色调量化,井依据混沌置乱算子,将色调结果置乱,然后构造平均误差最小的特征集合C,最后依据误差扩散数据隐藏算法将置乱后水印图像隐藏于原始图像中;在认证端,从接收到的图像提取其中所隐藏水印信息并进行逆置乱,比较接收到的图像和反置乱后的隐藏信息,判断内容发生变化的位置,并依据所提取的水印信息修复被篡改图像。实验结果表明,该算法对删除、替换、篡改等破坏图像内容的恶意操作有精确的检测和定位,以及自修复能力。     

Image annotation using high order statistics in non-Euclidean spaces

Automatic image annotation is a promising way to achieve more effective image retrieval and image analysis by using keywords associated to the image content. Due to the semantic gap between low-level visual features and high-level semantic concepts of an image, however, the performances of many existing algorithms are not so satisfactory. In this paper, a novel image classification scheme, named high order statistics based maximum a posterior (HOS-MAP), is proposed to deal with the issue of image annotation. To bridge the gap between human judgment and machine intelligence, the proposed scheme first constructs a dissimilarity representation for each image in a non-Euclidean space; then, the information of dissimilarity diffusion distribution for each image is achieved with respect to the high-order statistics of a triplet of nearest neighbor images; finally, a maximum a posteriori algorithm with the information of Gaussian Mixture Model and dissimilarity diffusion distribution is adopted to estimate the relevance between each annotation and an input un-annotated image. Experimental results on a general-purpose image database demonstrate the effectiveness and efficiency of the proposed automatic image annotation scheme.     

基于多级描述模型的渐进式图像内容理解

针对目前基于内容的图像检索技术中低级特征无法准确全面地描述高级语义的问题,本文提出了一种基于多级图像描述模型的渐进式图像内容理解.该图像描述模型在不同层次上对图像内容进行分析和提取,实现了图像内容的全方位描述,从底层向高层的过渡是渐进式的图像理解过程.特别是从视觉感知层到目标层,体现了图像低级特征与高级语义之间的过渡.本文给出了一种基于先验知识的上下文驱动的目标理解算法,实现了图像语义的提取.作为一个应用实例,本文给出了以上方法在基于内容的图像检索技术中的具体应用.     

Image alignment based perceptual image hash for content authentication

Perceptual image hash is an emerging technology that is closely related to many applications such as image content authentication, image forging detection, image similarity detection, and image retrieval. In this work, we propose an image alignment based perceptual image hash method, and a hash-based image forging detection and tampering localization method. In the proposed method, we introduce an image alignment process to provide a framework for image hash method to tolerate a wide range of geometric distortions. The image hash is generated by utilizing hybrid perceptual features that are extracted from global and local Zernike moments combining with DCT-based statistical features of the image. The proposed method can detect various image forging and compromised image regions. Furthermore, it has broad-spectrum robustness, including tolerating content-preserving manipulations and geometric distortion-resilient. Compared with state-of-the-art schemes, the proposed method provides satisfactory comprehensive performances in content-based image forging detection and tampering localization.     

基于拉普拉斯金字塔的红外光强与偏振图像融合

利用红外偏振信息(偏振度、偏振角)对目标进行成像,可以更好地抑制图像的背景噪声,提高信噪比。而且偏振信息相对于光强信息一般会蕴含更丰富的目标边缘轮廓信息。因此,提出一种将红外辐射光强图像和偏振度图像进行融合的算法。此方法首先对参与融合的每幅图像分别进行拉普拉斯金字塔分解,获得每层的分解图像;然后对分解后的每层图像采用不同的融合方法进行图像融合,获得每层融合图像,并对每层融合后的图像进行图像重构,得到最后的融合结果。多幅图像融合后的效果表明该方法能够增加图像的信息量,有利于场景感知和目标识别。     

Texture analysis-based multi-focus image fusion using a modified Pulse-Coupled Neural Network (PCNN)

Multi-focus image fusion is an effective method of information fusion that can take a series of source images and obtain a fused image where everything is in focus. In this paper, a multi-focus image fusion method based on image texture that adopts a modified Pulse-Coupled Neural Network (PCNN) approach is proposed. First, the texture of an image is obtained by means of image cartoon and texture decomposition. An ignition image is then acquired by inputting the image textures into a modified PCNN. Ignition images are compared to each other to obtain an initial decision map. A small object detection and bilateral filter is then applied to the initial decision map to reduce noise and enable smoother processing. Finally, the source images and decision map are used to produce the fused image. Experimental results demonstrate that the proposed method effectively preserves the source images information while delivering good image fusion performance.     

Shadow removal using bilateral filtering

In this paper, we propose a simple but effective shadow removal method using a single input image. We first derive a 2-D intrinsic image from a single RGB camera image based solely on colors, particularly chromaticity. We next present a method to recover a 3-D intrinsic image based on bilateral filtering and the 2-D intrinsic image. The luminance contrast in regions with similar surface reflectance due to geometry and illumination variances is effectively reduced in the derived 3-D intrinsic image, while the contrast in regions with different surface reflectance is preserved. However, the intrinsic image contains incorrect luminance values. To obtain the correct luminance, we decompose the input RGB image and the intrinsic image. Each image is decomposed into a base layer and a detail layer. We obtain a shadow-free image by combining the base layer from the input RGB image and the detail layer from the intrinsic image such that the details of the intrinsic image are transferred to the input RGB image from which the correct luminance values can be obtained. Unlike previous methods, the presented technique is fully automatic and does not require shadow detection.