基于目标识别的红外与微光图像融合方法

 为了在融合图像中突出运动目标,提出了一种基于动态目标检测和识别的图像融合算法。先对红外图像序列中的运动目标进行检测和提取,同时对红外和微光图像进行融合,最后将提取到的红外目标与融合图像进行二次融合。试验结果表明,该算法获得的融合图像不仅具有普通融合算法信息丰富的特点,还具有鲜明的红外目标指示特性。     

基于目标识别的多传感器图像信息融合技术

本文针对目标识别问题，论述了多传感器目标属性融合技术。包括目标属性融合概念、原理、处理结构、算法及应用举例。指出了多传感器目标属性融合中存在的主要问题及解决方法，并预测了发展趋势     

基于级联神经网络的实用型三维复合不变性多目标识别

以三 飞机模型作为待识别目标,模型真实场景,对用于多目标分类识别的级联神经网络重新进行了研究。实验发现畜产品上降的主要原因是实际采集的目标发生的复杂畸变与计算机模拟产生的效果并不一样。用采集得到的目标图像作为训练样本,对网络重新构造和训练,取得了好的实验结果。分析了其中涉及到目标定位、图像分割等图像预处理问题,提出了一种基于二值图像开矿学腐蚀运算的快速目标检测位法,可快速有效地对目标进行检测定 。     

基于快速骨架提取的三维目标识别

针对三维目标识别问题,提出了一种基于快速骨架提取的方法。根据骨架所反映的目标拓扑结构,建立了不同目标局部结构之间的对应关系;而在对应的局部曲线段上,采用基于曲线配准的方法进行匹配;以各个局部匹配的成本之和评估不同目标的相似性.这种方法在目标出现一定程度的视觉变形时仍具有较好的识别效果,同时避免了基于曲线方法的匹配目标的局部,而忽略局部之间相互的空间组织的缺点所造成的误匹配.算例结果表明这种算法对于三维目标有较好的识别效果.     

用二维时域有限差分法分析条形多层波导

为精确分析条形多层波导和多量子阱波导，本文提出了一种以二维时 限差分法为基础的全波数值方法，给出了场分布的色散特性的数值结果，并与矢量有限元法已有的结果作了比较；文中还研究了以二维时域有限差分法为基础的标量近似技术，提出了它的有效性和精度的局限性。     

改进的对向传播网络及其在多传感器目标识别中的应用

针对多传感器数据融合和目标识别的特点,提出了改进的对向传播网络(MCPN),并与Dempster-Shafer(D-S)证据推理相结合,实现了决策层数据融合目标识别.文中利用仿真数据对所提出的网络训练算法和融合结构进行了实验研究.结果表明:改进后的对向传播网络识别性能优于传统的对向传播网络(CPN),融合后的目标识别率较单传感器明显提高.最后,将该方法应用于前视红外(FLIR)和可见光摄像机目标跟踪系统对算法和融合结构进行验证,结果表明文中提出的方法是可行的.     

基于多路可视图的健康与心梗患者心电图信号复杂网络识别

可视图(visibility graph, VG)算法已被证明是将时间序列转换为复杂网络的简单且高效的方法,其构成的复杂网络在拓扑结构中继承了原始时间序列的动力学特性.目前,单维时间序列的可视图分析已趋于成熟,但应用于复杂系统时,单变量往往无法描述系统的全局特征.本文提出一种新的多元时间序列分析方法,将心梗和健康人的12导联心电图(electrocardiograph, ECG)信号转换为多路可视图,以每个导联为一个节点,两个导联构成可视图的层间互信息为连边权重,将其映射到复杂网络.由于不同人群的全连通网络表现为完全相同的拓扑结构,无法唯一表征不同个体的动力学特征,根据层间互信息大小重构网络,提取权重度和加权聚类系数,实现对不同人群12导联ECG信号的识别.为判断序列长度对识别效果的影响,引入多尺度权重度分布熵.由于健康受试者拥有更高的平均权重度和平均加权聚类系数,其映射网络表现为更加规则的结构、更高的复杂性和连接性,可以与心梗患者进行区分,两个参数的识别准确率均达到93.3%.     

基于快速骨架提取的三维目标识别(英文)

针对三维目标识别问题,提出了一种基于快速骨架提取的方法。根据骨架所反映的目标拓扑结构,建立了不同目标局部结构之间的对应关系;而在对应的局部曲线段上,采用基于曲线配准的方法进行匹配;以各个局部匹配的成本之和评估不同目标的相似性.这种方法在目标出现一定程度的视觉变形时仍具有较好的识别效果,同时避免了基于曲线方法的匹配目标的局部,而忽略局部之间相互的空间组织的缺点所造成的误匹配.算例结果表明这种算法对于三维目标有较好的识别效果.     

基于数学形态谱和二维矢量分类网络的模式识别及二维矢量分类网络的…

介绍一种基于数学形态谱和二维矢量分类网络的模式识别体系。数学形态谱相对于图像平移和旋转不变，建立了光学二维矢量分类网络，利用光学逻辑操作和最大值网络的循环操作，得到与输入图像最佳匹配的模式。     

二维与三维氢原子的四类升、降算子

用因式分解法求出了二维和三维氢原子的角动量升、降算子，还利用合流超几何函数的递推关系求出了另外三类升、降算子．讨论了四类升、降算子的选择定则和守恒量子数．二维和三维氢原子有很大的相似性 关键词：     

Multi-Stream Convolution-Recurrent Neural Networks Based on Attention Mechanism Fusion for Speech Emotion Recognition

The quality of feature extraction plays a significant role in the performance of speech emotion recognition. In order to extract discriminative, affect-salient features from speech signals and then improve the performance of speech emotion recognition, in this paper, a multi-stream convolution-recurrent neural network based on attention mechanism (MSCRNN-A) is proposed. Firstly, a multi-stream sub-branches full convolution network (MSFCN) based on AlexNet is presented to limit the loss of emotional information. In MSFCN, sub-branches are added behind each pooling layer to retain the features of different resolutions, different features from which are fused by adding. Secondly, the MSFCN and Bi-LSTM network are combined to form a hybrid network to extract speech emotion features for the purpose of supplying the temporal structure information of emotional features. Finally, a feature fusion model based on a multi-head attention mechanism is developed to achieve the best fusion features. The proposed method uses an attention mechanism to calculate the contribution degree of different network features, and thereafter realizes the adaptive fusion of different network features by weighting different network features. Aiming to restrain the gradient divergence of the network, different network features and fusion features are connected through shortcut connection to obtain fusion features for recognition. The experimental results on three conventional SER corpora, CASIA, EMODB, and SAVEE, show that our proposed method significantly improves the network recognition performance, with a recognition rate superior to most of the existing state-of-the-art methods.     

基于神经网络的二维不变性目标识别方法

提出了一种基于多层前馈神经网络的二维不变性目标识别方法。利用傅里叶描述器提取具有旋转、平移及尺度不变性的目标形状特征。由于所识别的工业工具具有一个自由度,它们形状有一定的动态变化导致同一目标的形状特征敌意一的不唯一性。文中采用含有两个隐层的多层前馈网络学习及识别这些特征矢量。在实验中,在四类机械工具进行测试,并将所提出方法与最近邻分类器进行比较,结果表明,具有反向传输（ＢＰ）学习算法的多层前馈网络     

Fusion of 3D laser scanner and depth images for obstacle recognition in mobile applications

The problem of obstacle detection and recognition or, generally, scene mapping is one of the most investigated problems in computer vision, especially in mobile applications. In this paper a fused optical system using depth information with color images gathered from the Microsoft Kinect sensor and 3D laser range scanner data is proposed for obstacle detection and ground estimation in real-time mobile systems. The algorithm consists of feature extraction in the laser range images, processing of the depth information from the Kinect sensor, fusion of the sensor information, and classification of the data into two separate categories: road and obstacle. Exemplary results are presented and it is shown that fusion of information gathered from different sources increases the effectiveness of the obstacle detection in different scenarios, and it can be used successfully for road surface mapping.     

A Parallel Multi-Modal Factorized Bilinear Pooling Fusion Method Based on the Semi-Tensor Product for Emotion Recognition

Multi-modal fusion can exploit complementary information from various modalities and improve the accuracy of prediction or classification tasks. In this paper, we propose a parallel, multi-modal, factorized, bilinear pooling method based on a semi-tensor product (STP) for information fusion in emotion recognition. Initially, we apply the STP to factorize a high-dimensional weight matrix into two low-rank factor matrices without dimension matching constraints. Next, we project the multi-modal features to the low-dimensional matrices and perform multiplication based on the STP to capture the rich interactions between the features. Finally, we utilize an STP-pooling method to reduce the dimensionality to get the final features. This method can achieve the information fusion between modalities of different scales and dimensions and avoids data redundancy due to dimension matching. Experimental verification of the proposed method on the emotion-recognition task using the IEMOCAP and CMU-MOSI datasets showed a significant reduction in storage space and recognition time. The results also validate that the proposed method improves the performance and reduces both the training time and the number of parameters.     

Recognition of a light line pattern by Hu moments for 3-D reconstruction of a rotated object

A simple technique of a light line projection for 3-D shape detection of rotated objects is presented. In this technique, an object is rotated around its symmetrical axis four times at an angle by using an electromechanical device and scanned by a light line. Four views of the object surface are extracted from each one of these rotations by processing a set of light line images. These views are connected using rotation angle and origin coordinates to obtain the complete 3-D shape. Angle and origin are calculated by recognition of a light line pattern. Light line pattern is recognized by Hu moments. In this manner, measurement errors on setup are avoided. It is an advantage over common methods, where these two parameters are measured directly on the setup to obtain the 3-D shape. Local profilometric method is based on the perturbation that the light line suffers when it is projected on the object surface. This perturbation is observed on an image plane due to the different direction between light line projector and viewer. These perturbations are measured by using Gaussian functions. In this technique the light line images are processed in very fast form. The technique and processing time are presented in detail. This technique is tested with objects, which have little information and its experimental results are also presented.     

A Mahalanobis Surrogate-Assisted Ant Lion Optimization and Its Application in 3D Coverage of Wireless Sensor Networks

Metaheuristic algorithms are widely employed in modern engineering applications because they do not need to have the ability to study the objective function’s features. However, these algorithms may spend minutes to hours or even days to acquire one solution. This paper presents a novel efficient Mahalanobis sampling surrogate model assisting Ant Lion optimization algorithm to address this problem. For expensive calculation problems, the optimization effect goes even further by using MSAALO. This model includes three surrogate models: the global model, Mahalanobis sampling surrogate model, and local surrogate model. Mahalanobis distance can also exclude the interference correlations of variables. In the Mahalanobis distance sampling model, the distance between each ant and the others could be calculated. Additionally, the algorithm sorts the average length of all ants. Then, the algorithm selects some samples to train the model from these Mahalanobis distance samples. Seven benchmark functions with various characteristics are chosen to testify to the effectiveness of this algorithm. The validation results of seven benchmark functions demonstrate that the algorithm is more competitive than other algorithms. The simulation results based on different radii and nodes show that MSAALO improves the average coverage by 2.122% and 1.718%, respectively.     

A Macroscopically Relevant 3D‐Metrology Approach for Nanocatalysis Research

A 3D nanometrological approach, which considers as an unbiased validation criterion the quantitative match between values of properties determined by macroscopic characterization techniques and those determined from the nanoscopic results, is developed to unveil the details of complex nanocatalysts. This approach takes into account both the peculiar characteristics of this type of materials and the large influence of noise in the tilt series. It combines, in an optimized way, the latest experimental developments in high angle annular dark field scanning transmission electron microscopy mode (HAADF‐STEM) tomography, such as batch tomography, image denoising by undecimated wavelet transforms, improved reconstructions by total variation minimization and a more efficient, user‐independent, segmentation scheme. To illustrate the use of this novel approach, the 3D structural characterization of a model nanocatalyst comprising gold nanoparticles dispersed on the surface of CeO₂ nanocubes is performed, and the obtained results used to compute the values of different macroscopic chemical and textural properties. Comparison with values obtained by macroscopic characterization techniques match very closely those obtained by 3D nanometrology. Importantly, the new approach described in this work also illustrates a pipeline for nearly fully automated HAADF‐STEM tomography studies, guaranteeing reliable correlations between nanoscopic and macroscopic properties.     

在并行网络上实现叶轮机械内部流动两类流面准三元迭代并行计算

叶轮机械S1／S2两类流面迭代计算具有天然可并行性，在科学与工程计算国家重点实验室的SGI工作站网络上首次实现了两类流面准三元迭代并行计算．     

A comparison of 2D and 3D digital image correlation for a membrane under inflation

Three-dimensional (3D) digital image correlation (DIC) is becoming widely used to characterize the behavior of structures undergoing 3D deformations. However, the use of 3D-DIC can be challenging under certain conditions, such as high magnification, and therefore small depth of field, or a highly controlled environment with limited access for two-angled cameras. The purpose of this study is to compare 2D-DIC and 3D-DIC for the same inflation experiment and evaluate whether 2D-DIC can be used when conditions discourage the use of a stereo-vision system. A latex membrane was inflated vertically to 5.41 kPa (reference pressure), then to 7.87 kPa (deformed pressure). A two-camera stereo-vision system acquired top-down images of the membrane, while a single camera system simultaneously recorded images of the membrane in profile. 2D-DIC and 3D-DIC were used to calculate horizontal (in the membrane plane) and vertical (out of the membrane plane) displacements, and meridional strain. Under static conditions, the baseline uncertainty in horizontal displacement and strain were smaller for 3D-DIC than 2D-DIC. However, the opposite was observed for the vertical displacement, for which 2D-DIC had a smaller baseline uncertainty. The baseline absolute error in vertical displacement and strain were similar for both DIC methods, but it was larger for 2D-DIC than 3D-DIC for the horizontal displacement. Under inflation, the variability in the measurements were larger than under static conditions for both DIC methods. 2D-DIC showed a smaller variability in displacements than 3D-DIC, especially for the vertical displacement, but a similar strain uncertainty. The absolute difference in the average displacements and strain between 3D-DIC and 2D-DIC were in the range of the 3D-DIC variability. Those findings suggest that 2D-DIC might be used as an alternative to 3D-DIC to study the inflation response of materials under certain conditions.     

基于MRI和深度学习的桥小脑角区脑膜瘤与听神经瘤分类算法研究

桥小脑角区脑膜瘤与听神经瘤是两种常见的脑部肿瘤,它们的临床表现和影像学表现极为相似,在临床诊断时极易发生误诊．将影像数据与深度学习方法相结合,建立脑膜瘤与听神经瘤的判别模型,可以为两种脑肿瘤的及时准确诊断提供重要手段．本文采集了307名脑肿瘤患者的T₁W-SE序列图像,通过对原始图像进行限制对比度自适应直方图均衡化（Contrast Limited Adaptive Histogram Equalization,CLAHE）等预处理,提升数据集图像质量,再经过建立的三维卷积神经网络（3-Dimensional Convolutional Neural Network,3D CNN）深度学习框架中图像特征的学习,实现对脑膜瘤与听神经瘤的分类．图像增强参数与网络结构参数经过优化后,对脑膜瘤与听神经瘤分类的准确率达到0.918 0,曲线下面积（Area Under Curve,AUC）为0.913 4,实现了对桥小脑角区脑膜瘤与听神经瘤的有效判别．