改进布鸟搜索算法最大熵值的医学图像分割

研究医学图像分割问题.医学图像是医学影像的分析基础,医学图像由于组织边缘模糊和灰度不均匀含噪声等特点,导致最大熵值分割医学图像算法难以进行准确分割,分割精度低,为了提高医学图像分割的准确性,提出一种改进布鸟搜索算法优化最大熵值的医学图像分割方法.首先由最大熵法找到医学图像分割目标函数,然后采用改进布谷鸟搜索算法对目标函数进行优化,找到医学图像的最佳分割点,实现医学图像分割,最后采用多幅医学图像进行仿真,以测试算法性能.结果表明,改进方法不仅解决了传统最大熵值医学图像分割算法存在的缺陷,同时提高医学图像分割的精度,并且具有较好的鲁棒性,具有较好的实际应用价值.     

一种改进的势函数聚类多阈值图像分割算法

针对基于势函数聚类的多阈值图像分割算法的不足,定义了伪势的概念,并在原算法基础上提出了一种改进的图像分割算法。由伪势概念确定了伪势合并的判别方法,按照此方法,当相邻的两个峰之间的距离小于所定义的自适应模糊伪势因子时,则应该进行伪势合并。改进后的算法在计算剩余势函数时判断是否存在伪势,然后在势划分函数组的确定过程中相应地进行伪势合并计算。利用多幅图像进行了多阈值分割的仿真试验,结果表明,改进的基于势函数的多阈值图像分割算法具有更好的鲁棒性和分割效果。     

基于改进飞蛾扑火算法的多阈值彩色图像分割

针对彩色图像多阈值分割存在计算量大、运行时间长等问题,在飞蛾扑火算法(Moth-Flame Optimization,MFO)的基础上,引入莱维飞行策略和自适应权重变化策略,提出LSMFO算法(Levy Self-adaptive Moth Flame Optimization)对最佳分割阈值进行优化搜索。为了验证该算法的有效性,选取4幅伯克利大学经典图像,将LSMFO算法与另外5种元启发式算法进行对比。应用Otsu方法进行多阈值图像分割实验,并用SSIM、PSNR、EPI三个指标评估分割后的图像效果。实验结果显示,LSMFO算法在指标衡量比较上整体水平优于其他算法,表明该算法运行时间短、分割精度高,能够有效解决彩色图像多阈值分割问题。     

基于数据流和精确定位的多线程行人探测系统

行人防撞警告系统(Pedestrian Collision Warning System,PCWS)是车载主动安全系统的一项主要功能.目前通常的行人检测算法在商用化过程中的主要制约是庞大的计算量导致的低处理帧率.研究了PCWS中的关键技术,综合运用了基于VP评估的空间限制滑动窗口分布、基于数据流的多线程处理流程、基于CENTRIST特征的快速线性SVM分类器、基于直方图交叉核SVM(HIKSVM)的非线性分类器等多种加速技术,达到了实时处理的要求.对于CENTRIST特征不能精确描述对象轮廓所造成的定位不准问题,通过基于高斯权重分布的极大化抑制以及基于外轮廓边缘差异微调包围框尺寸来精确重定位探测框,以满足商用化对测距准确性的要求.     

改进的多阈值动态二值化算法

在模式识别中,二值化效果的好坏直接影响着识别率。文章在研究现有二值化算法的基础上,对多阈值动态二值化进行了改进,并以身份证为例来验证该算法的有效性,实验结果表明所提算法有较高的抗噪音能力。     

时间序列复杂性自动度量：快速的弹簧测试

提出了时间序列复杂性自动度量的弹簧测试的一种提速方法。弹簧测试包括5阈值压缩空间中的弹簧图,以及该弹簧图的自相似性刻画数值CC。针对其速度较慢的问题,应用峰谷值映射简化弹簧图,提速自相似性度量。测试方程使用小波映射混沌、统一方程、Rossler超混沌方程。以分岔图为基本的对比判别几何特征,以基于FFT的谱熵复杂度为自动度量金标准,对比改进前后弹簧自相似性度量CC值的结果,给出各测试例的稳定性分析及计算时间比较。结果表明：改进后的弹簧测试能明显区别周期与混沌,增加了超混沌与混沌的辨识度,时间上提升2-20倍。本工作为周期、混沌、随机三分类识别的自动选参的信号复杂性度量贡献了改良的弹簧测试方法,本质是快速提取三维相图（弹簧图）的非对称非线性统计不变量。     

Interference probability matrix for disassembly sequence planning under uncertain interference

Disassembly of end-of-life products is a common step in remanufacturing and recycling. Disassembly sequence planning is the process that automatically finds the optimal sequence of components being removed. A key element of disassembly sequence planning is a suitable mathematical representation that describes the interference of any two components in a product. Previous studies on disassembly sequence planning have tended to focused on the interference that is fixed and known. However, the interference may be uncertain due to complex end-of-life conditions such as deformation, corrosion and rust. To deal with uncertain interference, this paper proposes an interference probability matrix as a new mathematical representation that uses probability to indicate uncertainty in the interference, and establishes a multi-threshold planning scheme to generate the optimal disassembly sequence plans. Three case studies are given to demonstrate the use of the proposed approach. It is also tested the performance of four multi-objective optimization algorithms that can be adopted in the proposed multi-threshold planning scheme.     

Fast image segmentation based on multi-resolution analysis and wavelets

An efficient algorithm for image segmentation based on a multi-resolution application of a wavelets transform and feature distribution is presented. The original feature space is transformed into a lower resolution with a wavelets transform to derive fast computation of the optimum threshold value in a feature space. Based on this lower resolution version of the given feature space, a single feature value or multiple feature values are determined as the optimum threshold values. The optimum feature values, which are in the lower resolution, are projected onto the original feature space. In this step a refinement procedure may be added to detect the optimum threshold value. Experimental results for the proposed algorithm indicate feasibility and reliability for fast image segmentation.     

基于混沌电磁场优化算法的多阈值彩色图像分割

为了解决彩色图像多阈值分割中计算时间长、分割精度低的问题,在电磁场优化算法(Electromagnetic Field Optimization,EFO)的基础上引入一种混沌策略用于算法初始化中,提出混沌电磁场优化算法(Chaotic Electromagnetic Field Optimization,CEFO)对图像的最佳阈值向量进行搜索。将其与另外5种优化算法进行对比,采用PSNR、MSSIM和FSIM 3个图像质量评价指标和算法运行时间(CPU Time)对6种分割算法进行分析比较。结果表明,CEFO具有收敛速度快、分割精度高的优势,能够胜任多阈值彩色图像分割的工程任务。     

Using multi-threshold threshold gates in RTD-based logic design: A case study

The basic building blocks for resonant tunneling diode (RTD) logic circuits are threshold gates (TGs) instead of the conventional Boolean gates (AND, OR, NAND, NOR) due to the fact that, when designing with RTDs, TGs can be implemented as efficiently as conventional ones, but realize more complex functions. Recently, RTD structures implementing multi-threshold threshold gates (MTTGs) have been proposed which further increase the functionality of the original TGs while maintaining their operating principle and allowing also the implementation of nanopipelining at the gate level. This paper describes the design of n-bit adders using these MTTGs. A comparison with a design based on TGs is carried out showing advantages in terms of power consumption and power delay product.