Line diffusion: a parallel error diffusion algorithm for digital halftoning

Floyd and Steinberg's (1976) error diffusion technique is a well-known approach to digital halftoning. The main drawback of this technique is that it is inherently serial. This paper presents a new parallelizable error-diffusion algorithm, calledline diffusion. In this method, the pixels of the original image are divided into classes line by line, and all the pixels on a line are halftoned simultaneously. Errors are distributed randomly. Experimental results show that line diffusion is comparable to error diffusion in image quality. A sequential line diffusion algorithm is also provided.     

数字半色调技术中的误差扩散算法的研究

误差扩散是一种非常优秀的调频半色调技术,但这种方法在图像高光和暗调以及阶调平滑区域会产生“蠕虫”,同时噪声较多,导致半色调图像的边缘不清晰。在Floyd和Steinberg的误差扩散算法的基础上,设计一种新的基于平均阈值和改变扫描路径的误差扩散算法——阈值采用平均值法,通过优化平均阈值方法中的γ值得到最优的半色调图像,同时扫描路径采用“蛇型”路径。实验结果表明该算法可以减少“蠕虫”,增强图像的边缘,提高半色调图像的主观视觉效果。     

Confined‐error‐diffusion algorithm for flat‐panel display

Abstract— The reduction of a structural pattern at specific gray levels caused by digital halftone methods is the subject of this paper. This problem is more severe in some flat‐panel displays because their black levels typically are brighter than other display blocks. A patented halftone algorithm, confined error diffusion (CED), that confines the error‐carry within the dither mask is described and extended. First, the CED algorithm that dynamically applies random error diffusion or the ordered‐dither method, depending upon image content, is described in detail. Finally, we propose an advanced CED algorithm for improving the gradation characteristics of the CED algorithm. The performance of the proposed algorithms is compared to the experimental results for natural test images. In order to verify the halftone quality, a structural similarity measure for color images by taking into account the interrelation between color channels is proposed, and the results based on the proposed method, the color similarity measure method, is given.     

有序抖动半调图像压缩算法

结合有序抖动半调图像特性,针对原有矢量量化初始码书算法存在非典型码字、运算量大和空间分布不均匀等欠缺,提出一种能在训练集中均匀分布的初始码书生成方法,应用到LBG算法中,并结合无损压缩方法,进行二次压缩。实验表明,本方法压缩比高、实现简单,且解码后图像视觉效果得到明显改善。     

基于调色板分析及误差扩散抖动的迷彩仿造

仿造迷彩是针对特定的环境背景生成的伪装迷彩。在迷彩纹理设计上,无规则的、点阵式的数码迷彩相对传统迷彩具有更好的迷惑效果。提出一种新的仿造数码迷彩方法,在调色板颜色分类排序及直方图合理分析的基础上提取四种主色,结合抖动半色调方法统计四主色比例,并运用随机像素点集的方式生成迷彩图案。实验证明,该方法生成的迷彩图与参照背景图能自然融合,达到较高的伪装隐蔽效果。     

多尺度框架误差分散逆半调方法研究

摘 要：通过分析现有基于滤波的误差分散逆半调方法,提出了一种多尺度逆半调框架。基于该框架,引入平稳小波变换,提出了一种新的逆半调算法。首先,对半调图进行3尺度离散平稳小波变换,通过分析各尺度图像的频谱特性,构造了HIST滤波器,并结合尺度间小波系数的相关性,采用不同策略分别抑制各尺度图像中的噪声,最后利用小波反变换得到逆半调图像。实验验证了多尺度逆半调框架的正确性,并且该算法所得逆半调图像具有良好的视觉效果,能保持图像边缘,峰值信噪比也有所提高。     

抖动半调图像跳黑块无损压缩算法

提出了一种有序抖动半调图像跳黑块无损压缩方法。通过分析抖动半调图像特性,对半调图进行分块及块间异或预处理,使其转变为较大面积黑色区域中夹杂着零星白点的二值图像,接着运用跳黑块编码法对其无损压缩。在跳黑块编码中,对只有一个白像素的非黑块采用特定的短码字编码;对其余类型的非黑块采用直接编码。实验结果表明,改进的跳黑块编码能在一定程度上对非黑块进行有效压缩,且新算法不但具有较高的压缩效率,时间、空间复杂度也较低。     

Robustness design of fuzzy controllers for nonlinear multiple time-delay singularly perturbed systems: using dither as auxiliary

The stabilisation problem is considered in this study for nonlinear multiple time-delay singularly perturbed (NDSP) systems. First, a neural-network (NN) model is employed to approximate the reduced-order model of the NDSP plant. Then, a linear differential inclusion (LDI) state-space representation is established for the dynamics of the NN model. Based on this LDI state-space representation, a robustness design of fuzzy control is proposed to stabilise the NDSP system. If the designed fuzzy controller cannot stabilise the NDSP system, a dither (as an auxiliary of the fuzzy controller) is simultaneously introduced to stabilise the NDSP system. If the frequency of dither is high enough, the outputs of the dithered system and its corresponding mathematical model, the relaxed system, can be made as close as desired. This makes it possible to obtain a rigorous prediction of the stability of the dithered system by establishing that of the relaxed system. Finally, this study provides a numerical example with simulations to illustrate the concepts discussed throughout this article.     

Multiscale hydro-thermo-mechanical model for early-age and mature concrete structures

Temperature and early-age mechanical properties in hydrating concrete structures present a significant risk for cracking, having a major impact on concrete durability. In order to tackle these phenomena, a multiscale analysis is formulated. It accounts for a high variety of cement properties, concrete composition, structure geometry and boundary conditions.The analysis consists of two steps. The first step focuses on the evolution of moisture and temperature fields. An affinity hydration model accompanied with non-stationary heat and moisture balance equations are employed.The second step contains quasi-static creep, plasticity and damage models. It imports the previously calculated moisture and temperature fields into the mechanical problem in the form of a staggered solution. The whole model has been implemented in the ATENA software, including also the effect of early-age creep, autogenous and drying shrinkage.Validation on selected structures shows a good prediction of temperature fields during concrete hardening and a reasonable performance of the mechanical part.     

An efficient query learning algorithm for ordered binary decision diagrams

In this paper, we propose a new algorithm that exactly learns ordered binary decision diagrams (OBDDs) with a given variable ordering via equivalence and membership queries. Our algorithm uses at most n equivalence queries and at most 2n (log₂ m + 3n) membership queries, where n is the number of nodes in the target-reduced OBDD and m is the number of variables. The upper bound on the number of membership queries is smaller by a factor of O(m) compared with that for the previous best known algorithm proposed by [R. Gavaldà, D. Guijarro, Learning Ordered Binary Decision Diagrams, Proceedings of the 6th International Workshop on Algorithmic Learning Theory, 1995, pp. 228–238].     

A linear time algorithm for computing the convex hull of an ordered crossing polygon

In this paper, we describe a linear time algorithm for finding the convex hull of an ordered crossing polygon and prove its correctness.     

Development of high-accuracy image centroiding algorithm for CMOS-based digital sun sensors

A digital sun sensor calculates the incident sunlight angle using the sunlight image registered on a Complimentary Metal Oxide Semiconductor (CMOS) image sensor. In order to accomplish this, an exact center of the sunlight image has to be determined. Therefore, an accurate estimate of the centroid is the most important factor in digital sun sensor development. The most general method for determining the centroid is the thresholding method, and this method is also the simplest and easy to implement. Another centering algorithm often used is the image filtering method that utilizes image processing. The sun sensor accuracy using these methods, however, is quite susceptible to noise in the detected sunlight intensity. This is especially true in the thresholding method where the accuracy changes according to the threshold level. In this paper, a template method that uses the sunlight image model to determine the centroid of the sunlight image is suggested, and the performance has been compared and analyzed. The template method suggested, unlike the thresholding and image filtering method, has comparatively higher accuracy. In addition, it has the advantage of having consistent level of accuracy regardless of the noise level, which results in a higher reliability.     

Bees algorithm for design of dual‐beam linear antenna arrays with digital attenuators and digital phase shifters

This article describes a method of designing a reconfigurable dual‐beam linear antenna array using bees algorithm (BA). The BA is an optimization algorithm inspired by the behavior of the honey bees to find the optimal way of harvesting food resources around the hive. The proposed method is very simple and can be used directly in practice to synthesize multiple beam antenna arrays with digital attenuators and digital phase shifters. A good agreement between the desired pattern and the synthesized pattern using BA is obtained. © 2008 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2008.     

基于多分辨率的PET图像优质有序子集最大期望重建算法

在正电子发射断层扫描(PET)成像中,最大似然期望(MLEM)算法因不能有效地抑制噪声、收敛速度慢而难以直接应用于临床。有序子集最大期望(OSEM)算法具有较快的收敛速度,但是在迭代一定次数之后重建质量会迅速下降。针对此问题,将多分辨率技术引入到有序子集最大期望重建算法的子集中,以此抑制噪声,同时稳定求解过程。实验结果表明,新的重建算法克服了传统算法图像退化的缺点,并具有加快算法收敛速度的优点,能获得较高的信噪比(SNR)和较好的图像视觉效果。     

四元声传感器面阵快速测向算法及误差分析

为了对低速移动的声源目标进行较高实时性的方向跟踪,把四元声传感器面阵分为两个二元线阵,分别对目标进行测向,然后融合两组测量结果,给出了目标方位角的计算公式.根据线阵在被跟踪目标方位角不同时测向精度的不同,用加权因子增加较高精度的线阵测量结果的影响,降低误差较大的线阵测量结果的影响.该算法可对传感器阵列所在平面内任意方位的声源快速进行测向,完成一次测向只需对两对时间采样序列分别进行互相关运算,并可获得较高的定向精度,通过实验验证了该算法的精度和有效性.     

Sequential phase feeding algorithm with minimal beam pointing error for centrosymmetric planar array antenna

Phased array antennas using digital phase shifters for beam steering are characterized by beam pointing error. This study proposes a sequential phase feeding algorithm (SPFA) with minimal beam pointing error for a centrosymmetric planar array antenna. The theoretical expression of a beam pointing error caused by phase quantification error is derived. On the basis of the derivation results, the phase shift of each element is sequentially quantified through truncating or carrying to achieve a smaller beam pointing error. The proposed algorithm improves the beam pointing accuracy of the centrosymmetric planar array antenna, and the computational load is acceptable. Simulation results verify the effectiveness of SPFA.     

A quantized water cycle optimization algorithm for antenna array synthesis by using digital phase shifters

In this article, a quantized water cycle algorithm (QWCA) is used for the antenna array pattern synthesis with low side‐lobe levels (SLLs) and nulls at desired directions by using four‐bit digital phase shifters. In addition to the standard features as a metaheuristic algorithm, QWCA has an internal quantization mechanism and a precalculated array factor method. The latter provides an accelerated procedure to QWCA under favour of the digitized values that can be stored in a three‐dimensional array. This acceleration is based on the reality that the accessing data in the memory need less time than the usage of the mathematical functions throughout the optimization process. The internal quantization mechanism of QWCA is utilized to achieve digital values matching to the discrete values of the phase shifter instead of the simple rounding up/down routines after optimization. The numerical results showed that QWCA can obtain very good SLLs and null depth levels (NDLs) on the synthesized pattern. Moreover, the results are achieved in remarkably short optimization times. SLL and NDL values obtained by QWCA are also compared with the available literature values. The comparisons reveal that QWCA is able to produce better results than the other compared alternatives. © 2014 Wiley Periodicals, Inc. Int J RF and Microwave CAE 25:21–29, 2015.     

Time second-order finite difference/finite element algorithm for nonlinear time-fractional diffusion problem with fourth-order derivative term

In this article, we study and analyze a Galerkin mixed finite element (MFE) method combined with time second-order discrete scheme for solving nonlinear time fractional diffusion equation with fourth-order derivative term. We firstly introduce an auxiliary variable $\sigma =△u$, reduce the fourth-order problem into a coupled system with two equations, discretize the obtained coupled system at time $t_{k?\frac{\alpha }{2}}$ by a second-order difference scheme with second-order approximation for fractional derivative, then formulate mixed weak formulation and fully discrete MFE scheme. Further, we give the detailed proof for stability of scheme, the existence and uniqueness of MFE solution, and a priori error estimates. Finally, by some numerical computations, we test the theoretical results, which illustrate that we can obtain the numerical results for two variables, moreover, we arrive at second-order time convergence orders, which are higher than the ones yielded by the $L1$-approximation.