A new hybrid approach for dynamic continuous optimization problems

A new hybrid approach for dynamic optimization problems with continuous search spaces is presented. The proposed approach hybridizes efficient features of the particle swarm optimization in tracking dynamic changes with a new evolutionary procedure. In the proposed dynamic hybrid PSO (DHPSO) algorithm, the swarm size is varied in a self-regulatory manner. Inspired from the microbial life, the particles can reproduce infants and the old ones die. The infants are especially reproduced by high potential particles and located near the local optimum points, using the quadratic interpolation method. The algorithm is adapted to perform in continuous search spaces, utilizing continuous movement of the particles and using Euclidian norm to define the neighborhood in the reproduction procedure. The performance of the new proposed approach is tested against various benchmark problems and compared with those of some other heuristic optimization algorithms. In this regard, different types of dynamic environments including periodic, linear and random changes are taken with different performance metrics such as real-time error, offline performance and offline error. The results indicate a desirable better efficiency of the new algorithm over the existing ones.     

Polynomial approximation of functions of matrices and applications

In solving a mathematical problem numerically, we frequently need to operate on a vector by an operator that can be expressed asf(A), whereA is anN ×N matrix [e.g., exp(A), sin(A), A^–-]. Except for very simple matrices, it is impractical to construct the matrixf (A) explicitly. Usually an approximation to it is used. This paper develops an algorithm based upon a polynomial approximation tof (A). First the problem is reduced to a problem of approximatingf (z) by a polynomial in z, where z belongs to a domainD in the complex plane that includes all the eigenvalues ofA. This approximation problem is treated by interpolatingf (z) in a certain set of points that is known to have some maximal properties. The approximation thus achieved is almost best. Implementing the algorithm to some practical problems is described. Since a solution to a linear systemAx=b isx=A ^–1 b, an iterative solution algorithm can be based upon a polynomial approximation tof (A)=A ^–1. We give special attention to this important problem.     

Amino acid profiles and liposomes: Their role as chemosensory information carriers in the marine environment

Yellowfin tuna enhance their hunting success in the vast pelagic environment by using their sense of smell to detect intact (uninjured) prey that are beyond visual range. However, the olfactory cues that tuna use would normally face huge and rapid dilution in the open ocean. We demonstrate that these prey odors are complexed within biologically derived lipid structures that probably delay the dilution of the amino acids to subthreshold concentrations and provide persistent arousal and search cues for the tuna. This may be the first demonstration of an extracorporeal biological function for liposomes. Tuna may also form chemical search images to maximize feeding efficiency. We demonstrate that the amino acid profiles of various prey species are consistent over time and between schools, which makes the formation of search images feasible.     

基于CMOS工艺的无线图像传感器网络综述

无线传感器网络是无线通讯技术、网络技术、传感器技术交叉融合的新兴产物。无线传感器网络具备小型化、智能化、网络化和能够无线通讯等特点,能够满足人们在生产生活、科学研究、军事、医疗等众多领域对于实时信息获取的需求。由于人类获取信息的80%以上是通过视觉获取的,因此,无线图像传感器网络作为人类获取视觉图像信息的重要工具在无线传感器网络中占有重要的地位。无线图像传感器网络要想实现小型化、智能化、集成化等要求,采用与标准CMOS工艺兼容的CMOS图像传感器技术将是最为理想的选择。随着CMOS图像传感器工艺和设计水平的进步,CMOS图像传感器在成像质量、生产成本、小型化、智能化等方面显示出巨大的优势,同时由于与VLSI电路兼容的天然特性,基于CMOS图像传感器的无线图像传感器网络拥有巨大的发展潜力和广阔的市场前景。基于CMOS的无线图像传感器网络目前呈现多元化的发展趋势,高分辨率、高速、低功耗、智能化等方向是基于CMOS的无线图像传感器网络未来发展的方向。     

改进的形态学运算在声纳图像生成中的应用

声纳技术应用日益广泛,已成为海洋测量的重要工具,而去除噪声处理是对声纳图像进行正确判读的前提,本文采用了改进的形态学算法减少声纳图像中噪声的影响,对图像进行预处理,采用Matlab图形图像处理工具对处理的算法进行仿真的方式,从而使处理效果达到最佳。     

SVD域的图像高斯噪声强度估计

精确估计图像或视频中的噪声强度对于后续的信号处理是至关重要的先决条件。通过对含噪图像的奇异值特性的研究,提出一种精确的SVD域的图像噪声强度估计算法。该算法对噪声强度估计提出了创新的解决方法:1)利用奇异值的尾部数据进行噪声强度估计,这样达到尽可能地降低图像信息对噪声估计的干扰;2)对含噪图像加入已知强度的高斯白噪声,以计算噪声估计时需要设置的与图像内容相关的参数,因此该算法可以自适应图像的结构,能够广泛地适应各种类型的图片。实验结果表明SVD域噪声强度估计算法适用于各种图片类型,而且在极大的噪声强度范围内都能够稳定精确地估计噪声强度。     

基于卡尔曼滤波改进的精子图像序列分割方法

图像分割是精子图像识别的一项关键技术,在精子运动能力分析中起着至关重要的作用。本文对采集的连续精子图像序列进行灰度化、去噪等预处理后,采用Otsu算法对首幅动物精子图像二值化,对后续图像采用Kalman Filter确定二值化阈值范围,改进Otsu算法求出每一幅图像的适当阈值并进行二值化,缩短算法时间并能保证分割精度。应用形态学消除精子尾部和部分精子之间的粘连现象,通过计算和比较目标面积、形状因子,去除小颗粒杂质以及形状及灰度和精子相似的杂质,为精子运动能力检测提供高质量的分割图像。     

关于Photoshop插件的SAR图像处理探究

SAR图像处理主要包括相干斑抑制、特征提取等,如果在使用Photoshop插件SAR图像处理图片,将主要精力集中于算法的设计和优化方面,可以尽量减少在用户界面、内存管理和基本的图像操作方面的编程,既可以实现多种图像处理,又容易实现功能的充分扩大与填充,从而在最短的时间内做做多的事。     

Multi-Scale Dilated Convolutional Neural Network for Hyperspectral Image Classification

基于多尺度空洞卷积神经网络的高光谱图像分类

郑姗姗¹,刘文¹,单锐¹,赵静一²,江国乾³,张智⁴

（1. 燕山大学理学院,河北 秦皇岛 066004;2. 燕山大学机械工程学院,河北 秦皇岛 066004;3. 燕山大学电气工程学院,河北 秦皇岛 066004;4. 北京航天研究所,北京 100094）

创新点说明：

1）将图像分割方法——空洞卷积用于卷积神经网络进行高光谱图像分类,提取更加广泛、抽象的图像特征。

2）构建基于多尺度空洞卷积神经网络的高光谱图像分类方法。搭建多尺度聚合结构,在每个通道中使用快捷连接和空洞卷积结构,有效提取图像特征,避免信息丢失。

研究目的：

针对图像信息丢失问题,得到高精度的高光谱图像分类方法。

研究方法：

在Indian Pines和Pavia University数据集上,与4个已有的高光谱图像分类方法进行对比实验,比较OA, AA和Kappa值。

研究结果：

多尺度空洞卷积神经网络在Indian Pines和Pavia University数据集上OA值分别达到了99.58%,99.92%。AA值分别达到了99.57%,99.90%。Kappa分别达到了99.52%,99.89%。

结论：

1）在卷积神经网络中引入空洞卷积,可以有效避免图像信息丢失。

2）多尺度空洞卷积神经网络能提取更佳的判别性特征,实现高分类性能。

关键词：多尺度聚合;空洞卷积;高光谱图像分类;快捷连接

     

深度学习的人体图像半自动标注系统

针对目前数据标注过于依赖硬件、手动数据标注效率低下的问题,提出了基于深度学习的人体图像半自动标注系统.系统通过对算法进行改进,增加人体关键点个数进行特征提取和加入运动信息的约束,提高了视频分阶段标注的准确率.使用真实数据集仿真实验证明了通过深度学习算法进行数据标注的可行性,并且使用半自动标注的速度快、准确率高.