A novel hardware-oriented Kohonen SOM image compression algorithm and its FPGA implementation

Kohonen self-organizing map (K-SOM) has proved to be suitable for lossy compression of digital images. The major drawback of the software implementation of this technique is its very computational intensive task. Fortunately, the structure is fairly easy to convert into hardware processing units executing in parallel. The resulting hardware system, however, consumes much of a microchip’s internal resources, i.e. slice registers and look-up table units. This results in utilising more than a single microchip to realize the structure in pure hardware implementation. Previously proposed K-SOM realizations were mainly targetted on implementing on an application specific integrated circuit (ASIC) with low restriction on resource utilization. In this paper, we propose an alternative architecture of K-SOM suitable for moderate density FPGAs with acceptable image quality and frame rate. In addition, its hardware architecture and synthesis results are presented. The proposed K-SOM algorithm compromises between the image quality, the frame rate throughput, the FPGA’s resource utilization and, additionally, the topological relationship among neural cells within the network. The architecture has been proved to be successfully synthesized on a single moderate resource FPGA with acceptable image quality and frame rate.     

基于改进Census变换的自由曲面立体匹配方法

针对基于网格投影的自由曲面三维重建方法中,特征点集中分布于投影网格线条之上,而投影网格线条未能覆盖的部分即投影网格线条间的自由曲面区域无特征点,导致重建精度较低的问题,提出一种基于自适应窗口的改进Census变换立体匹配方法,对投影网格线条间的自由曲面进行三维重建;将特征匹配算法与区域匹配算法相结合,以图像中目标物体上的投影网格指导Census变换;利用可变视差范围计算Hamming距离作为匹配代价;通过筛选获得精度较高的匹配点对。实验结果表明,该算法应用于双目视觉图像三维重建,可有效提高投影网格线条间自由曲面的重建精度。     

一种稀疏度自适应的稀疏傅里叶变换算法

稀疏快速傅里叶变换需要信号以傅氏域的稀疏度为先验信息,但稀疏度通常是未知的,在一定程度上限制了算法的应用。为此,提出一种新的稀疏傅里叶变换算法。在下采样域进行能量检测,得到稀疏度的初始值,通过增大下采样维度提高稀疏度估计的准确性,从而近似估计稀疏度,设定阈值剔除冗余信息从而得到较好效果。实验结果表明,当信号长度大于219或稀疏度小于900时,该算法性能优于西方快速傅里叶变换,且具有较强的鲁棒性。     

基于改进Census变换的多特性立体匹配算法

针对当前立体匹配算法存在的匹配准确率低，难以达到实用的高精度水平的问题，提出了一种基于改良的Census变换与色彩信息和梯度测度相结合的多特性立体匹配算法，实现高精度的双目立体匹配。算法首先在初始代价匹配阶段，将改进的Census变换、色彩和梯度测度赋权求和得出可靠的初始匹配代价；在聚合阶段，采取高效快捷的最小生成树聚合，获得匹配代价矩阵；最后根据胜者为王法则得到初始视差图，并引入左右一致性检测等策略优化视差图，获得高精度的视差图，实验阶段对源自Middlebury上的标准测试图进行测试验证，实验结果表明，经本文算法处理得到的15组测试数据集的视差图在非遮挡区域的平均误匹配率为6.81%，算法实时响应性优良。     

GAKREM: A novel hybrid clustering algorithm

We introduce a novel clustering algorithm named GAKREM (Genetic Algorithm K-means Logarithmic Regression Expectation Maximization) that combines the best characteristics of the K-means and EM algorithms but avoids their weaknesses such as the need to specify a priori the number of clusters, termination in local optima, and lengthy computations. To achieve these goals, genetic algorithms for estimating parameters and initializing starting points for the EM are used first. Second, the log-likelihood of each configuration of parameters and the number of clusters resulting from the EM is used as the fitness value for each chromosome in the population. The novelty of GAKREM is that in each evolving generation it efficiently approximates the log-likelihood for each chromosome using logarithmic regression instead of running the conventional EM algorithm until its convergence. Another novelty is the use of K-means to initially assign data points to clusters. The algorithm is evaluated by comparing its performance with the conventional EM algorithm, the K-means algorithm, and the likelihood cross-validation technique on several datasets.     

A prism-based single-lens stereovision system: From trinocular to multi-ocular

This paper investigates a passive stereovision system which employs one single CCD camera and one pyramid-like glass prism. Its trinocular variety is first presented: each image captured by this system can be split into three sub-images and these sub-images are taken as the images simultaneously captured by three virtual cameras which are generated by the prism. Two different approaches are developed to model this system: one bases on a conventional camera calibration technique and the other bases on geometrical analysis of ray sketching. The second approach is a relatively simpler but is sufficiently accurate as compared to the calibration based approach. Then the knowledge on this trinocular system is extended to build a single-lens multi-ocular stereovision system. Experiments are conducted to validate this system. The ideas presented in this paper are believed novel.     

基于小波的自适应图像数字盲水印算法

该文提出一种在小波变换域自适应嵌入图像数字盲水印的新算法,在水印检测过程中不需要原始图像。首先根据三级小波的细节系数确定嵌入位置,利用系数间存在很强的相关性,在第二级小波的水平和垂直细节数据中,通过采用修改图像小波系数及其邻域系数的均值之间的大小关系来自适应的嵌入置乱后的水印,嵌入的强度由该处的边缘强弱确定,这样,能在水印嵌入的鲁棒性和不可见性方面达到很好的平衡。并且,小波嵌入的位置互不相邻,嵌入水印信息之间不会产生相互干扰,从而提高水印提取的质量。该算法实验结果表明,该数字水印算法能很好的抵抗附加噪声、裁减、有损压缩、滤波等各种攻击,具有较强的不可见性和鲁棒性。     

Data Adaptive Linear Decomposition Transform

Cadzow, J. A., and Yammen, S., Data Adaptive Linear Decomposition Transform, Digital Signal Processing12 (2002) 494–523In this paper a novel method for decomposing one-dimensional data sequences is developed. It is of the wavelet variety with important distinctions. A standard wavelet transform entails down-sampling the responses of two fixed structured filters to the sequence being decomposed to produce the half length detail sequence and a coarse sequence. The new transform entails processing down-sampled data by an optimum interpolation filter which seeks to approximate the odd indexed sequence elements by a linear combination of neighboring even indexed elements. The resulting interpolation error sequence of half length plays the role of the wavelet detail sequence. Unlike a traditional wavelet transform, the interpolation filter employed is adapted to the data being analyzed. This data dependency typically produces improved performance relative to traditional wavelet transforms. This enhanced performance is demonstrated on a number of standard test signals. Furthermore, this new transform is computationally efficient due to the inherent parallel structure of the data decomposition method.     

Adaptive differential evolution algorithm with novel mutation strategies in multiple sub-populations

Differential evolution (DE) algorithm has been shown to be a very effective and efficient approach for solving global numerical optimization problems, which attracts a great attention of scientific researchers. Generally, most of DE algorithms only evolve one population by using certain kind of DE operators. However, as observed in nature, the working efficiency can be improved by using the concept of work specialization, in which the entire group should be divided into several sub-groups that are responsible for different tasks according to their capabilities. Inspired by this phenomenon, a novel adaptive multiple sub-populations based DE algorithm is designed in this paper, named MPADE, in which the parent population is split into three sub-populations based on the fitness values and then three novel DE strategies are respectively performed to take on the responsibility for either exploitation or exploration. Furthermore, a simple yet effective adaptive approach is designed for parameter adjustment in the three DE strategies and a replacement strategy is put forward to fully exploit the useful information from the trial vectors and target vectors, which enhance the optimization performance. In order to validate the effectiveness of MPADE, it is tested on 55 benchmark functions and 15 real world problems. When compared with other DE variants, MPADE performs better in most of benchmark problems and real-world problems. Moreover, the impacts of the MPADE components and their parameter sensitivity are also analyzed experimentally.     

基于自适应权重的改进Census立体匹配算法

针对传统Census算法对噪声敏感且在弱纹理区域匹配精度低的不足,提出一种基于自适应权重的改进算法.在代价计算阶段,通过空间相似度加权计算得到参考像素值,设定阈值限定参考值与中心点像素的差异,使算法能够判断中心点是否发生突变并自适应选择中心参考像素值.在代价聚合阶段,引入多尺度聚合策略,将引导滤波作为代价聚合核函数,加...     

一种有效的Gk-prototypes聚类算法

针对传统的聚类算法对初始聚类中心敏感、只能对单一属性聚类且聚类效果有时欠佳等不足,提出了一种能处理数值属性和分类属性的Gk-prototypes聚类算法。在经典的k-prototypes聚类算法的基础上,利用去模糊相似矩阵来构造粗粒子集,结合粒计算和最大最小距离法确定初始聚类中心,并改进了目标函数。实验结果和理论分析表明,Gk-prototypes聚类算法与其他基于k-prototypes的改进算法相比,聚类更准确,有效性更好,鲁棒性更强。     

一种新颖的改进萤火虫算法

为了克服萤火虫算法收敛速度慢和易于早熟的缺陷,本文提出了一种新颖的改进混沌萤火虫算法(Improved Chaos Firefly Algorithm,ICFA)。ICFA算法利用逻辑映射混沌序列设计了一种混沌局部搜索算子,试图提升算法的收敛速度;ICFA算法利用立方映射混沌序列设计了一种混沌替换算子,试图避免算法的早熟收敛:同时,本文将ICFA算法与目前最有竞争力的改进萤火虫算法进行了一系列的比较研究。6个高维多峰函数的测试结果表明,与其他高效的改进萤火虫算法相比,本文提出的混沌萤火虫算法具有收敛速度快、寻优能力强的优点,是目前最优秀的改进萤火虫算法之一。     

A novel complex-valued bat algorithm

Bat algorithm is a recent optimization algorithm with quick convergence, but its population diversity can be limited in some applications. This paper presents a new bat algorithm based on complex-valued encoding where the real part and the imaginary part will be updated separately. This approach can increase the diversity of the population and expands the dimensions for denoting. The simulation results of fourteen benchmark test functions show that the proposed algorithm is effective and feasible. Compared to the real-valued bat algorithm or particle swarm optimization, the proposed algorithm can get high precision and can almost reach the theoretical value.     

A novel stochastic optimization algorithm

This paper presents a new stochastic approach SAGACIA based on proper integration of simulated annealing algorithm (SAA), genetic algorithm (GA), and chemotaxis algorithm (CA) for solving complex optimization problems. SAGACIA combines the advantages of SAA, GA, and CA together. It has the following features: (1) it is not the simple mix of SAA, GA, and CA; (2) it works from a population; (3) it can be easily used to solve optimization problems either with continuous variables or with discrete variables, and it does not need coding and decoding,; and (4) it can easily escape from local minima and converge quickly. Good solutions can be obtained in a very short time. The search process of SAGACIA can be explained with Markov chains. In this paper, it is proved that SAGACIA has the property of global asymptotical convergence. SAGACIA has been applied to solve such problems as scheduling, the training of artificial neural networks, and the optimizing of complex functions. In all the test cases, the performance of SAGACIA is better than that of SAA, GA, and CA.     

QSobel:A novel quantum image edge extraction algorithm

Edge extraction is an indispensable task in digital image processing. With the sharp increase in the image data, real-time problem has become a limitation of the state of the art of edge extraction algorithms.In this paper, QSobel, a novel quantum image edge extraction algorithm is designed based on the flexible representation of quantum image(FRQI) and the famous edge extraction algorithm Sobel. Because FRQI utilizes the superposition state of qubit sequence to store all the pixels of an image, QSobel can calculate the Sobel gradients of the image intensity of all the pixels simultaneously. It is the main reason that QSobel can extract edges quite fast. Through designing and analyzing the quantum circuit of QSobel, we demonstrate that QSobel can extract edges in the computational complexity of O(n2) for a FRQI quantum image with a size of2 n × 2n. Compared with all the classical edge extraction algorithms and the existing quantum edge extraction algorithms, QSobel can utilize quantum parallel computation to reach a significant and exponential speedup.Hence, QSobel would resolve the real-time problem of image edge extraction.     

A novel Minkowski-distance-based consensus clustering algorithm

Consensus clustering is the problem of coordinating clustering information about the same data set coming from different runs of the same algorithm. Consensus clustering is becoming a state-of-the-art approach in an increasing number of applications. However, determining the optimal cluster number is still an open problem. In this paper, we propose a novel consensus clustering algorithm that is based on the Minkowski distance. Fusing with the Newman greedy algorithm in complex networks, the proposed clustering algorithm can automatically set the number of clusters. It is less sensitive to noise and can integrate solutions from multiple samples of data or attributes for processing data in the processing industry. A numerical simulation is also given to demonstrate the effectiveness of the proposed algorithm. Finally, this consensus clustering algorithm is applied to a froth flotation process.     

一类自适应混沌文化算法

混沌变异进化算法忽略了混沌规律性,未充分利用知识来提高算法的局部收敛能力.为此,借鉴文化算法的双层进化结构,在文化算法的进化引导函数中引入自适应混沌变异策略,提出一种自适应混沌文化算法.利用进化过程隐含知识控制变异尺度,使知识引导个体能跳出局部较优解,在保证种群多样性的同时,实现进化后期的精细搜索.仿真结果表明,该算法可以有效提高进化收敛速度,具有较好的计算稳定性.