基于背景估计和集成分类的眼底硬性渗出检测

眼底图像中硬性渗出的检测对于糖尿病性视网膜病变(简称"糖网")的早期诊断具有重要意义.为了实现眼底图像中硬性渗出物的自动检测,本文提出了一种结合背景估计和集成分类的眼底图像硬性渗出物自动检测方法.首先,对图像进行亮度校正、去噪、对比度增强等预处理操作,然后结合形态学方法进行背景估计和图像重建并移除视盘区域,得到渗出物候选区域.最后利用集成分类方法对候选区域进行分类,提取最终的硬性渗出区域.实验结果表明,本方法能够有效准确地检测到眼底图像中的硬性渗出物,对于糖网自动诊断技术的研究具有积极意义.     

融合重建滤波器和暗特征的渗出物检测算法

针对多数视网膜渗出物提取算法检测精度不高的 问题,提出了一种融合重建滤波器和 暗特征的视网 膜渗出物检测算法。首先,对视网膜灰度图像均值滤波后应用“球形”重建滤波器移除视网 膜血管,使用列 滤波器标识出渗出物区域并阈值分割得到含有视盘和边界的渗出物区域图像。其次,在均值 滤波后的灰度 图像中检测视盘和图像边界,并根据视盘构建掩膜,接着从含有视盘和边界的渗出物区域图 像中移除视盘 和边界,然后应用 “圆盘形”重建滤波器扩充渗出物尺寸。最后,对视网膜灰度图像自适 应直方图均衡化, 阈值分割后反转灰度值得到暗特征,从扩充渗出物尺寸后的图像中移除暗特征即得到最终渗 出物图像。实验结果表明,本文算法能有效检测视网膜眼底图像中的渗出物,较现有算法具 有更高的检测精度。     

基于多模型融合和区域迭代生长的视网膜血管自动分割

糖尿病视网膜病变是成年人致盲首因,视网膜血管分割是诊断糖尿病视网膜病变的基础.为提高视网膜血管分割准确性,提出一种基于多模型融合和区域迭代生长的视网膜血管自动分割算法.首先,预处理后分别构建数学形态学、匹配滤波器、尺度空间分析、多尺度线检测和神经网络模型初步分割视网膜血管,为减少噪声取五个分割结果的均值作为初步输出.其次,设计掩膜分离渗出物和视盘,将数学形态学模型分割结果替换掩膜白色区域,并融合初步输出生成组合结果.最后,考虑视网膜血管先验知识,对组合结果阈值分割和区域迭代生长后获取最终结果.实验结果表明,该算法分割DRIVE和STARE眼底图像库视网膜血管的检测精度、敏感度和特异性分别为0.9457、0.7843、0.9815以及0.9472、0.7826、0.9803,优于多数经典算法.     

基于深度学习的糖尿病眼底病变检测研究

糖尿病不仅会增加视网膜血管疾病的风险，严重时甚至会发展成为糖尿病视网膜病变。糖尿病视网膜病变的4种典型病理特征是微动脉瘤、出血、硬性渗出物和软性渗出物。随着机器学习尤其是深度学习的发展，智能辅助诊断医疗已经成为一种趋势，智能辅助诊断的前提是可以定性定量地提取出相应的病变区域。提出了一种基于深度学习级联架构参数优化的眼底病变检测模型，该模型有效解决了眼底病变的多尺度和小目标问题，在DDR数据集上检测病变的综合测试精度达0.380，检测性能优于目前主流的检测网络。     

眼底图像中黄斑中心与视盘自动检测新方法

在眼底图像自动分析中,视盘与黄斑的定位是糖尿病性视网膜病变计算机辅助诊断或筛查的先决条件。该文提出一种应用方向局部对比度滤波结合局部血管密度的方法,直接先行检测黄斑中心再行定位视盘,不同于现有的先行检测视盘或血管再行定位黄斑的一般方法,有效地提高黄斑定位正确率,能更好地应用于糖尿病性黄斑水肿的自动评估。实验选取了网络公开的HEI-MED数据集中169幅黄斑水肿眼底图像,黄斑和视盘的定位正确率同步达到98.2%,算法简单且无监督,优于现有的方法,具有良好的临床应用前景。     

基于可控图像分割的快速视网膜血管提取算法

针对多数视网膜血管提取算法实时性不强和分割 精度不高的问题,提出了一种基于可控图像分割的 快速视网膜血管提取算法。首先,对视网膜G分量图像的灰 度进行反转和自适应直方图均衡化,应用结 构元素为“菱形”和“圆盘形”的形态学“开”运算平滑图像背景和增强血管对比度,消除 视盘后阈值分割并二值 化得到不含视盘的分割图像。其次,根据在灰度图像中检测到的视盘构建掩膜,再次对 视网膜绿色分 量图像自适应直方图均衡化后进行阈值分割,并和掩膜进行逻辑“与”运算得到含有掩膜的 分割图像。最后, 将不含视盘的分割图像与含有掩膜的分割图像进行逻辑“与”运算,并融合边界信息获得最 终的视网膜血管 结构。实验结果表明,本文算法能有效提取视网膜眼底图像的血管网络,有较强的实时性和 较高的分割精度。     

糖尿病视网膜病变眼底图像筛查研究

眼底照相是获取眼部图像的主要技术之一。利用眼底相机对视网膜病变区域进行拍摄可以获得清晰的图像,从获取的图像中能够直接观察到眼球中的渗出物、出血点和微血管瘤,根据检测出的病灶类型、数量和位置等信息可进行糖尿病视网膜病变分类。基于此,本文利用深度神经网络对糖尿病视网膜病变进行自动分类识别,提出了一种体系结构简单、在通用设备上运行速度快的卷积神经网络CA-RepVGG(CA代表Channel Attention, RepVGG为现有模块)。利用单路极简结构的RepVGG模块替代复杂的可使用性较差的模块作为分类模型的主体部位,并选用高效通道注意力机制ECA替代压缩注意力机制SE,以此来提升模型对病变分级的能力。此外,本文还将CA-RepVGG模型与传统的分类模型VGG-16、Inception-V3、ResNet-50和ResNext-50模型进行了比较。从比较结果可以看出,虽然CA-RepVGG模型的参数量最大,但由于其是单分支结构,且只有3×3卷积块,因此它的模型复杂度并不高,分类速度很快,比另外4个模型中分类速度最快的ResNet-50还高出15.3%。另外,利用两个混淆矩阵展示了所提模...     

多尺度特征融合双U型视网膜分割算法

视网膜血管形态结构是反映人体健康的重要指标 ,针对现有视网膜血管分割存在主 血管模糊、微细血管断裂和视盘误分割等问题,提出多尺度特征融合双U型视网膜分割算 法。首先,利用低层U-Net高效循环残差模块对眼底图像进行粗粒度分割,得到视网膜血 管 初步轮廓。其次,将粗分割图与原始特征图像素相乘送入高层U-Net,利用其缩放宽残差 模 块进行细粒度图像解码,丰富视网膜血管细节信息。同时利用3路径注意力机制复合性连接 双网络的编码层与解码层,实现特征映射跨网络传播,减小上下文语义差异。最后,融合双 层网络输出提取血管区域,双U 型网络能够更深层次提取血管像素,精准分割出视网膜细 节。在DRIVE与STARE数据集上进行实验,其准确率分别为96.45%和97.02%,敏感度分 别为83.35%和81.40%,特异性分别为98.38%和 98.83%,总体性能优于现有算法。     

基于CTCNet对球结膜进行糖尿病视网膜病变分类

糖尿病视网膜病变(diabetic retinopathy, DR)是一种糖尿病性微血管病变,会在球结膜微血管上有所体现,球结膜血管图像的获取比眼底图像更加便捷,但微血管的特征变化微小且难以量化。为了能够对患者进行早期辅助诊断,本文依据球结膜微血管形态与DR的关联,首先对球结膜图像进行预处理,使用限制对比度自适应直方图均衡(contrast limited adaptive histogram equalization, CLAHE)算法进行图像增强,随机处理使数据增强,然后结合卷积神经网络(convolutional neural network, CNN)和Transformer各自的网络优势构建CTCNet,对处理后的球结膜血管图像进行DR分类,分类准确率达到了97.44%,敏感度97.69%,特异性97.11%,精确度97.69%,通过实验对比CNN和Transformer, CTCNet网络性能优于其他模型,能够有效识别DR。     

基于局部傅里叶变换的方向自适应匹配滤波的视网膜血管分割

眼底图像中视网膜血管的结构对眼底疾病的分析和诊断具有重要的意义。针对匹配滤波方法中滤波参数的选取问题,文章提出了一种基于局部傅里叶变换的方向自适应匹配滤波的视网膜血管分割算法。文章算法首先对眼底图像进行预处理;通过分析预处理后的眼底图像中的局部傅里叶变换的能量分布,提取出血管点的主方向,同时利用Harris角点检测方法来校正血管分叉点的主方向;然后对每个像素点进行对应方向的匹配滤波;接着利用可变阈值方法对滤波后的图像进行分割;最后使用面积阈值法消除非血管点和噪声等。文章提出的算法对国际上公开的DRIVE库和STARE库进行了测试,实验结果证明了文中算法能够自适应地获取匹配滤波时的角度。     

Multisensor towed array detection system for UXO detection

The multisensor towed array detection system (MTADS) was designed to be an efficient, sensitive tool for the detection and characterization of buried unexploded ordnance. It comprises arrays of total-field magnetometers and time-domain electromagnetic induction (EMI) sensors, associated navigation and data acquisition hardware, and a custom data analysis system. The MTADS has conducted eleven demonstrations and surveys. The system has shown the ability to detect ordnance at its likely self-penetration depths with a probability of detection of 0.95 or better The model-derived positions and depths of the detected ordnance items are generally well within the physical size of the targets, making remediation much quicker and less costly than with standard techniques. Data sets corresponding to many of the MTADS surveys are available to others in the field     

Robust noise detection for speech detection and enhancement

A simple and robust method of reliably detecting stationary noise periods in a mobile telephony environment is presented. Such noise detection techniques are useful for updating parameters in adaptive voice activity detectors and for speech enhancement techniques, such as spectral subtraction, which require accurate noise models     

3D saliency detection based on background detection

Unlike 2D saliency detection, 3D saliency detection can consider the effects of depth and binocular parallax. In this paper, we propose a 3D saliency detection approach based on background detection via depth information. With the aid of the synergism between a color image and the corresponding depth map, our approach can detect the distant background and surfaces with gradual changes in depth. We then use the detected background to predict the potential characteristics of the background regions that are occluded by foreground objects through polynomial fitting; this step imitates the human imagination/envisioning process. Finally, a saliency map is obtained based on the contrast between the foreground objects and the potential background. We compare our approach with 14 state-of-the-art saliency detection methods on three publicly available databases. The proposed model demonstrates good performance and succeeds in detecting and removing backgrounds and surfaces of gradually varying depth on all tested databases.     

TD-CDMA系统中的联合检测技术--联合检测算法

多用户检测（MUD）技术作为CDMA系统的关键技术之一，具有优良的抗多址干扰（MAI）和抗远近效应性能。在介绍传统的多用户检测算法的基础上，引入了线性联合检测器，深入分析和比较了TD-CDMA系统中的联合检测算法，并根据TD-CDMA系统上行链路传播模型，对TD-CDMA上行链路的联合检测算法进行仿真。最后，根据仿真结果，比较了三种线性联合检测算法的性能。     

被动探测中的信号检测技术研究

被动探测中,低信噪比下检测信号的方法备受关注。本文讨论基于总和的检测方法。采用该方法就能将信号和噪声在时域中区分开来,并能较精确地确定信号的起止时刻。该方法具有实时性好、检测概率高的特点。仿真结果表明这种方法切实可行。     

Nonparametric detection

This paper considers some of the simpler nonparametric detection schemes and compares their asymptotic relative efficiencies to those of detectors which are optimal in the Neyman-Pearson sense. In the one-input case, the nonparametric sign and Wilcoxon detectors are compared to the linear detector which is optimal for the detection of a dc signal of unknown amplitude in Gaussian noise. For two-input systems the nonparametric polarity coincidence correlator is compared to the system which is optimal for the detection of a common random Gaussian component in two-input Gaussian noises. The nonparametric detectors are shown to offer advantages in simplicity of implementation and in insensitivity to changes in input statistics while performing moderately well compared to the parametric detectors. More impressive results can be obtained with more complicated detectors utilizing nonlinear rank statistics.     

Image forgery detection algorithm based on U-shaped detection network

Aiming at the defects of traditional image tampering detection algorithm relying on single image attribute,low applicability and current high time-complexity detection algorithm based on deep learning,an U-shaped detection network image forgery detection algorithm was proposed.Firstly,the multi-stage feature information in the image by using the continuous convolution layers and the max-pooling layers was extracted by U-shaped detection network,and then the obtained feature information to the resolution of the input image through the upsampling operation was restored.At the same time,in order to ensure higher detection accuracy while extracting high-level semantic information of the image,the output features of each stage in U-shaped detection network would be merged with the corresponding output features through the upsampling layer.Further the hidden feature information between tampered and un-tampered regions in the image upon the characteristics of the general network was explored by U-shaped detection network,which could be realized quickly by using its end-to-end network structure and extracting the attributes of strong correlation information among image contexts that could ensure high-precision detection results.Finally,the conditional random field was used to optimize the output of the U-shaped detection network to obtain a more exact detection results.The experimental results show that the proposed algorithm outperforms those traditional forgery detection algorithms based on single image attribute and the current deep learning-based detection algorithm,and has good robustness.     

认知无线电能量检测与循环谱检测算法研究

频谱检测是认知无线电的关键技术之一.通过对认知无线电频谱检测技术中的能量检测和循环谱检测技术进行研究,并对能量检测性能、观测数据长度和接收信噪比之间的定量关系进行分析;同时,还介绍了循环谱检测算法的原理,并仿真了不同信噪比下能量检测和循环谱检测算法的性能.仿真结果表明,循环谱检测算法具有比能量检测算法更高的检测灵敏度.     

激光主动侦测系统探测能力分析

激光主动侦测系统可实现远距离猫眼目标的成像探测与识别,探测能力分析是系统设计的前提和基础.为了评估激光主动侦测系统对猫眼目标的探测能力,以激光回波功率为基础,建立了系统信噪比的数学物理模型.分析了影响探测能力的因素,数值模拟了信噪比与接收光学镜头直径、发射激光峰值功率、激光束散角以及等效反射面离焦量的定量关系.结果表明,在一定作用距离范围内,散粒噪声、背景噪声与暗电流噪声、热噪声相差至少2个数量级,系统噪声主要来源于散粒噪声及背景噪声;提高系统探测能力最为有效的方法是选用激光束散角小的光源.这一结果可用于指导激光主动侦测系统的设计.     

Sequential detection versus conventional detection: A comparative study

M-ary sequential detection algorithms are discussed in terms of the advantage measure A, the number of possible signals M, and the signal-to-noise ratio SNR. The advantage measure A is defined as the logarithm of the ratio of the error probability of the optimum conventional detector to that of the optimum sequential detector with the same values of M and SNR. It increases as the SNR increases but decreases when M increases. For M → ∞, the advantage measure A diminishes to zero, so that application of the sequential approach is useless for large values of M. Thus, the sequential approach is most useful for binary detection with high signal-to-noise ratios. These results are demonstrated for detection of equiprobable orthogonal signals received in white Gaussian noise.