基于非负矩阵分解的图像脆弱水印算法

提出一种基于非负矩阵分解NMF(Non-negative Matrix Factorization)的脆弱数字水印算法。算法利用用户密钥构造NMF基矩阵,并在图像NMF分解过程中保持不变,二值水印图像嵌入NMF分解系数矩阵。实验结果本算法具有较强的鲁棒性,同时用户密钥保证的算法的脆弱性。     

基于约束NMF的盲源分离算法

非负矩阵分解(NMF)是解决非独立源信号混合的盲分离的另一条新途径.该文提出一种基于约束NMF的盲源分离算法,在对NMF估计得到的源信号施加最小相关约束的基础上,对混合矩阵估计施加行列式约束,实现NMF的唯一分解.与已有算法相比,本算法放宽了对混合矩阵的稀疏性要求,大幅提高了信号分离质量.该算法仍适用于独立源信号分离问题.     

基于约束随机分块的NMF图像哈希算法

基于非负矩阵分解(Non-negative Matrix Factorization, NMF)的图像哈希(image hashing)算法对图像有损压缩,低通滤波、尺度拉伸等处理具有很好的稳健性,但对图像旋转比较敏感。为此,该文在对NMF哈希算法的分块模式进行深入研究的基础上,提出一种可抗旋转攻击的NMF图像哈希算法。该方法通过对随机分块的区域进行限制,并选择合适的分块尺寸来减轻旋转攻击对图像造成的不良影响,从而提高了特征的旋转稳健性。实验表明,所提出的图像哈希算法在保持原NMF哈希算法对其它攻击稳健性的同时,能有效地抵抗旋转攻击。     

NMF-NAD：基于NMF的全网络流量异常检测方法

提出了一种基于非负矩阵分解(NMF,non-negative matrix factorization)的多元异常检测算法(NMF-NAD,NMF based network-wide traffic anomalies detection),该算法首先采用非负子空间方法对流量矩阵进行重构,然后基于重构误差利用Shewhart控制图进行异常检测。模拟实验与因特网实测数据的分析表明,NMF-NAD算法有较高的检测精度和较低的处理复杂度。     

改进NMF优化的水声目标信号增强算法

水声目标的特征相似和海洋环境噪声的多变,使得非负矩阵分解(NMF)算法的信号增强效果不佳.为此,提出基于改进NMF的增强算法,该算法一方面使用实际海洋环境噪声实时更新背景噪声基矩阵,以增强基向量的匹配性,另一方面对特征基矩阵进行相似检测去冗余,以消除系数分散造成的基向量丢失,最后构建增强滤波器实现目标信号的重构增强.实验结果表明,相比于正则化NMF算法、改进正交匹配追踪方法,所提算法取得最优的信号增强效果,并通过实测数据轴频提取实验结果进一步验证了所提算法的有效性.     

一种应用于人脸识别的有监督NMF算法

为了提高非负矩阵分解(NMF)算法识别率,提出了一种有监督的NMF(SNMF)方法.该算法对NMF基图像进行判别分析,然后选择主要反应类内差异的基图像来构造子空间,最后在子空间上进行识别.通过UMIST人脸库和CMU PIE人脸库上的实验结果表明,该方法对光照、姿态和表情变化具有一定的鲁棒性,识别率高于NMF方法和其它子空间分析法.     

快速NMF盲源分离算法

将秩一非负矩阵分解应用于盲源分离问题,把基于欧式距离的目标函数转化成二次函数的形式;施加稀疏性约束和正交性约束保证信号可分离性;利用二次函数的性质分别推得混合矩阵和源信号的迭代公式,从而得到一种基于秩一分解的快速NMF盲源分离算法（NMF-R1)。分析得到一次迭代更新NMF-R1算法比传统NMF盲源分离算法（NMF-BM)所需乘法次数少约30%,NMF-R1算法无矩阵求逆运算,NMF-BM算法还需2次矩阵求逆运算。图像信号的超定和欠定盲源分离仿真结果表明,NMF-R1算法都能分离出源信号, NMF-BM算法只能分离超定混合信号;NMF-R1算法与NMF-BM算法比,分离性能好、收敛速度快。      

非负矩阵集分解

非负矩阵分解(Nonnegative Matrix Factorization,NMF)是一种新近被提出的方法,它以非线性的方式实现对非负多元数据的纯加性、局部化、线性和低维描述。NMF可使数据中的潜在结构、特征或模式变得清晰,因此它作为一种有效的特征提取手段已被成功应用在许多领域的研究中。但是,NMF 的处理对象本质上是向量,用NMF处理数据矩阵集时要先将被处理矩阵集中的矩阵逐一矢量化,这常使对应的学习问题成为典型的小样本问题,从而使NMF结果的描述力不强、推广性差。为克服这两个问题,并保留NMF的好的特性,该文提出了非负矩阵集分解(Nonnegative Matrix-Set Factorization,NMSF),不同于NMF处理数据矩阵的矢量化结果,NMSF直接处理数据矩阵本身。理论分析显示:处理数据矩阵集时,NMSF会比NMF描述力强、推广性好。为了说明NMSF如何实现,也为了能对NMSF的性能做实验验证,构造了NMSF实现方式之一的基于双线性型的NMSF(Bilinear Form-Based NMSF,BFBNMSF)算法。BFBNMSF和NMF的比较实验结果支持了理论分析的结论。需要指出,更佳的描述力和更好的推广性意味着NMSF比NMF更善于抓住数据矩阵的本质特征。     

基于Alpha-NMF的AD样本分类及特异性基因选择方法

由于基因表达谱数据的高噪声、高维性、高冗余以及数据分布不均匀等特点使得在分析过程中仍然有很多挑战性问题。基于该目的,将一种无监督学习方法--非负矩阵分解方法,应用到基因表达谱数据中,挖掘出与AD相关的信息基因。然而标准NMF算法其效率较低,并且在基因表达数据的应用有效性低。为了适应该领域的需求,采用了Alpha-NMF算法。该算法能够有效的克服标准NMF算法的缺陷,获得较好的实验结果。多次运行Alpha-NMF算法,选取分类准确率和稳定性最优的实验结果,对其集合基因设定一阈值,筛选出集合基因中大于该阈值的信息基因。最后通过基因功能分类以及生物功能结构图来验证所提炼出的特异性基因的有用性和可靠性。     

基于稀疏性非负矩阵分解的偏振图像快速融合方法

针对基于非负矩阵分解(non-negative matrix factorization, NMF)的偏振图像融合方法效率低的不足,提出一种基于稀疏性NMF的偏振图像快速融合方法。首先,以偏振信息解析得到的各偏振参量图像构造原始数据集,其次,对NMF增加稀疏性约束,利用稀疏表示下的在线字典学习算法进行快速分解,然后对分解得到的三幅特征基图像按清晰度和方差进行排序,将排序后的特征基图像经直方图匹配及HSI颜色映射后,变换到RGB颜色空间,得到融合图像。与基于NMF的方法相比,运行时间提高约120倍,达到约1.5 s完成一次融合过程。实验结果验证了该方法在改善融合效果的同时,运行效率明显提高。     

Nonnegative Matrix and Tensor Factorization

In these lecture notes, the authors have outlined several approaches to solve a NMF/NTF problem. The following main conclusions can be drawn: 1) Multiplicative algorithms are not necessary the best approaches for NMF, especially if data representations are not very redundant or sparse. 2) Much better performance can be achieved using the FP-ALS (especially for large-scale problems), IPC, and QN methods. 3) To achieve high performance it is quite important to use the multilayer structure with multistart initialization conditions. 4) To estimate physically meaningful nonnegative components it is often necessary to use some a priori knowledge and impose additional constraints or regularization terms (to control sparsity, boundness, continuity or smoothness of the estimated nonnegative components).     

Speed-up trend analysis for H.261 and model-based image coding algorithms using a parallel-pipeline model

Model-based and object-oriented coding algorithms are generally more computationally complex than current block-based image coding standards such as H.261, due primarily to the complexity of the image analysis they require. In this paper, simulations of H.261 and two model-based coding algorithms are analysed in terms of their computational complexity, and mapped onto a generalised image coder parallel-pipeline model. Example implementations of the H.261 coder and an object-oriented coder using general purpose parallel processor systems are then presented to confirm the validity of the performance trend analyses; these achieve maximum speedups of about 11 and 1.7, respectively, using up to 16 processors. The performance trends shown by the analysis indicate that the complexity of model-based coding algorithms, when combined with their reliance on coding picture differences and content-dependent algorithm execution times, interact to make it very difficult to achieve significant speed-up of sequential algorithms. Furthermore, the algorithm complexity and abstract data structures will make direct hardware implementations increasingly difficult. Overcoming these problems to achieve real-time model-based coders may require significant algorithmic compromises to be made.     

基于非负矩阵分解的多环芳烃成份识别

从重叠比较严重的混合物三维荧光光谱中恢复单一光谱信号,是光谱解析的难点。考虑到光谱内在的非负性,采用非负矩阵分解的投影梯度和交替最小二乘两种算法,并结合K均值初始化方法,来解析菲、芘、蒽3种芳烃混合物的三维荧光光谱数据,有效避免出现负数的分解结果,提取3种成份的三维荧光光谱,得到计算光谱与对应参考光谱的相似系数均大于0.970。计算结果表明,非负矩阵分解能够克服光谱重叠带来的干扰,有效提取光谱成份,从而实现对菲、芘、蒽的成份识别。其中,交替最小二乘的NMF算法更适合实时在线监测。     

改进投影梯度非负矩阵分解的单训练样本特征提取研究

人脸识别是当前人工智能和模式识别的研究热点。非负矩阵分解(NMF)能够反映样本的局部的内在的联系,可用于单样本特征提取,但时间复杂度较高。投影梯度(Projected Gradient,PG)优化方法大幅降低了NMF约束优化迭代问题的时间复杂度,但是单训练样本存在对本类信息量描述不足的缺点。为此,该文提出了一种基于改进的投影梯度非负矩阵分解 (Improved Projected Gradient Non-negative Matrix Factorization,IPGNMF) 的单训练样本特征提取方法。在进行PGNMF算子之前,先将训练样本作Gabor分解,分解后的Gabor子图像在各个方向上可以更加丰富的描述样本特征,最后将各个Gabor子图像的PGNMF特征进行融合,作为最终的识别特征。在对人脸库ORL,YEL与FERET的识别实验中,与经典的特征提取方法比较,证明了可以有效地解决单训练样本人脸识别的问题。     

Adaptive call admission control under quality of serviceconstraints: a reinforcement learning solution

We solve the adaptive call admission control (CAC) problem in multimedia networks via reinforcement learning (RL). The problem requires that network revenue be maximized while simultaneously meeting quality of service (QoS) constraints that forbid entry into certain states and use of certain actions. We show that RL provides a solution to this constrained semi-Markov decision problem and is able to earn significantly higher revenues than alternative heuristics. Unlike other model-based algorithms, RL does not require the explicit state transition models to solve the decision problems. This feature is very important if one considers large integrated service networks supporting a number of different service types, where the number of states is so large that model-based optimization algorithms are infeasible. Both packet-level and call-level QoS constraints are addressed, and both conservative and aggressive approaches to the QoS constraints are considered. Results are demonstrated on a single link and extended to routing on a multilink network     

基于边缘模型的数字图像后处理方法研究

基于边缘模型重建的数字图像后处理算法可以改善高倍压缩图像的主观和客观质量。文章通过对压缩图像退化原因以及现有的基于边缘模型的后处理算法的分析和讨论，提出了一种新的基于边缘模型的后处理算法及今后数字图像后处理算法的研究方向。     

Audio Source Separation Based on Residual Reprojection

This paper describes an audio source separation that is based on nonnegative matrix factorization (NMF) and expectation maximization (EM). For stable and high‐performance separation, an effective auxiliary source separation that extracts source residuals and reprojects them onto proper sources is proposed by taking into account an ambiguous region among sources and a source's refinement. Specifically, an additional NMF (model) is designed for the ambiguous region — whose elements are not easily represented by any existing or predefined NMFs of the sources. The residual signal can be extracted by inserting the aforementioned model into the NMF‐EM‐based audio separation. Then, it is refined by the weighted parameters of the separation and reprojected onto the separated sources. Experimental results demonstrate that the proposed scheme (outlined above) is more stable and outperforms existing algorithms by, on average, 4.4 dB in terms of the source distortion ratio.     

Model-based halftoning of color images

We present a new class of models for color printers. They form the basis for model-based techniques that exploit the characteristics of the printer and the human visual system to maximize the quality of the printed images. We present two model-based techniques, the modified error diffusion (MED) algorithm and the least-squares model-based (LSMB) algorithm. Both techniques are extensions of the gray-scale model-based techniques and produce images with high spatial resolution and visually pleasant textures. We also examine the use of printer models for designing blue-noise screens. The printer models cam account for a variety of printer characteristics. We propose a specific printer model that accounts for overlap between neighboring dots of ink and the spectral absorption properties of the inks. We show that when we assume a simple "one-minus-RGB" relationship between the red, green, and blue image specification and the corresponding cyan, magenta, and yellow inks, the algorithms are separable. Otherwise, the algorithms are not separable and the modified error diffusion may be unstable, The experimental results consider the separable algorithms that produce high-quality images for applications where the exact colorimetric reproduction of color is not necessary. They are computationally simple and robust to errors in color registration, but the colors are device dependent.     

In vivo fluorescence spectra unmixing and autofluorescence removal by sparse nonnegative matrix factorization

Fluorescence imaging locates fluorescent markers that specifically bind to targets; like tumors, markers are injected to a patient, optimally excited with near-infrared light, and located thanks to backward-emitted fluorescence analysis. To investigate thick and diffusive media, as the fluorescence signal decreases exponentially with the light travel distance, the autofluorescence of biological tissues comes to be a limiting factor. To remove autofluorescence and isolate specific fluorescence, a spectroscopic approach, based on nonnegative matrix factorization (NMF), is explored. To improve results on spatially sparse markers detection, we suggest a new constrained NMF algorithm that takes sparsity constraints into account. A comparative study between both algorithms is proposed on simulated and in vivo data.