Face Super-resolution Reconstruction and Recognition Using Non-local Similarity Dictionary Learning Based Algorithm

One of the challenges of face recognition in surveillance is the low resolution of face region. Therefore many superresolution (SR) face reconstruction methods are proposed to produce a high-resolution face image from one or a set of low-resolution face images. However, existing dictionary learning based algorithms are sensitive to noise and very time-consuming. In this paper, we define and prove the multi-scale linear combination consistency. In order to improve the performance of SR, we propose a novel SR face reconstruction method based on nonlocal similarity and multi-scale linear combination consistency (NLS-MLC). We further proposed a new recognition approach for very low resolution face images based on resolution scale invariant feature (RSIF). A series of experiments are conducted on two public face image databases to test feasibility of our proposed methods. Experimental results show that the proposed SR method is more robust and computationally effective in face hallucination, and the recognition accuracy of RSIF is higher than some state-of-art algorithms. © 2014 Chinese Association of Automation.

ISAR target recognition based on non-negative sparse coding

Aiming at technical difficulties in feature extraction for the inverse synthetic aperture radar （ISAR） target recognition, this paper imports the concept of visual perception and presents a novel method, which is based on the combination of non-negative sparse coding （NNSC） and linear discrimination optimization, to recognize targets in ISAR images. This method implements NNSC on the matrix constituted by the intensities of pixels in ISAR images for training, to obtain non-negative sparse bases which characterize sparse distribution of strong scattering centers. Then this paper chooses sparse bases via optimization criteria and calculates the corresponding non-negative sparse codes of both training and test images as the feature vectors, which are input into k neighbors classifier to realize recognition finally. The feasibility and robustness of the proposed method are proved by comparing with the template matching, principle component analysis （PCA） and non-negative matrix factorization （NMF） via simulations.

ON THE IDENTIFICATION OF RESIDUAL IMPACT PROPERTIES OF MATERIALS BY ACOUSTO-ULTRASONICS——A PATTERN RECOGNITION APPROACH

This paper deals with the application of Acousto-ultrasonics,in con- junction with Pattern Recognition and Classification techniques,to the identification of residual impact properties of a class of polymeric material,namely,Polyvinylchlo- ride(PVC).PVC specimens of different low-energy repeated impact damage states are processed by Acousto-ultrasonics(AU)to retrieve AU signals in the form of dig- italized records.These AU signals are grouped as distinct classes,each pertaining to a known level of repeated impact damage.Describing features of these AU signals are used to build Pattern Recognition(PR)Classifiers.These classifiers are used to identify unknown damage states in other PVC specimens by classifying the re- trieved AU signals as belonging to one of the classes.The obtained results indicate that Acousto-ultrasonics in combination with Pattern Recognition and Classification techniques can be used for the quantitative non-destructive identification of damage states in PVC specimens of unknown low-energy repeated impact conditions.

Recognition of Group Activities Using Complex Wavelet Domain Based Cayley-Klein Metric Learning

A group activity recognition algorithm is proposed to improve the recognition accuracy in video surveillance by using complex wavelet domain based Cayley-Klein metric learning.Non-sampled dual-tree complex wavelet packet transform(NS-DTCWPT)is used to decompose the human images in videos into multi-scale and multi-resolution.An improved local binary pattern(ILBP)and an inner-distance shape context(IDSC)combined with bag-of-words model is adopted to extract the decomposed high and low frequency coefficient features.The extracted coefficient features of the training samples are used to optimize Cayley-Klein metric matrix by solving a nonlinear optimization problem.The group activities in videos are recognized by using the method of feature extraction and Cayley-Klein metric learning.Experimental results on behave video set,group activity video set,and self-built video set show that the proposed algorithm has higher recognition accuracy than the existing algorithms.

肥胖儿童non-HDL-C、remnant cholesterol与亚临床动脉粥样硬化的关系

目的:了解non-HDL-C和remnant cholesterol与肥胖儿童亚临床动脉粥样硬化(AS)的关系。方法:根据甘油三酯(TG)水平将65例肥胖儿童(肥胖组)分为I组(TG≥2.3 mmol/L)32例,II组(TG<2.3 mmol/L)33例,与肥胖组儿童性别年龄相匹配的正常体质量儿童30例为III组(对照组)。比较3组儿童血脂指标,对肥胖组各项血脂指标与血浆致动脉粥样硬化指数(AIP)进行相关性分析,肥胖患儿体质指数、血脂指标与颈动脉内径、内膜-中层厚度(c-IMT)、血流参数进行相关性分析。结果:I组、II组的TG、总胆固醇(TC)、低密度脂蛋白胆固醇(LDL-C)、non-HDL-C、remnant cholesterol、AIP均较对照组升高,高密度脂蛋白胆固醇(HDL-C)较对照组降低,且I组的remnant cholesterol、AIP均较II组升高。肥胖组儿童AIP与TG、non-HDL-C、remnant cholesterol呈正相关,与HDL-C呈负相关。37例行颈动脉超声检查的肥胖儿童的c-IMT较正常儿童的增厚。结论:non-HDL-C、remnant cholesterol在肥胖儿童亚临床AS中有重要作用,在肥胖儿童血脂异常的治疗中应关注non-HDL-C、remnant cholesterol的变化。

Shenqi Fuzheng injection combined with GP chemotherapy in the treatment of advanced non-small cell lung cancer： a meta-analysis

Objective： To evaluate the clinical efficacy of Shenqi Fuzheng injection combined with gemcitabine plus cisplatin（GP） in the treatment of advanced non-small cell lung cancer （NSCLC）. Methods： we performed a systematicsearch in the electronic databases such as Cochrane Library, Pubmed, Embase, Chinese Journal Full-text Database,Chinese Biomedical Literature Database, Chinese Science and Technology Periodical Full-text Database andWanfang Database up to 30 January 2017. Randomized controlled trials （RCT） of Shenqi Fuzheng Injectioncombined with GP chemotherapy in the treatment of advanced NSCLC were searched, and all the RCTs wereconducted on methodological quality assessment. Data extraction and data analysis were according to standards ofCochrane systematic review. Results： Eight trials were included including a total of 701 patients. Meta-analysisresults： Shenqi Fuzheng injection combined with GP chemotherapy could significantly improve the functionalstatus of patients with NSCLC （OR = 3.44, 95% CI [2.26, 5.25], P 〈 0.0001） and clinical treatment efficacy （OR =（OR = 0.31, 95%CI [0.20, 0.47], P 〈 0.0001. The rate of leukopenia （OR = .31, 95%CI [0.20,0.47], P 〈 0.0001）,thrombocytopenia （OR = 0.58, 95%CI [0.37, 0.91], P = 0.020）, hemoglobin decline （（OR = 0.31, 95%CI [0.16,0.59], P = 0.0004） and incidence of gastrointestinal reactions （OR = 0.58,P 〈 0.05） could be reduced. Conclusion：Shenqi Fuzheng injection combined with GP chemotherapy in the treatment of advanced NSCLC obtainedsignificantly clinical efficacy. The quality of the literature incorporated is low, the conclusion requires high-qualityresearch to further prove.

Analysis of Numerical Results in High Temperature Congealment and Chemistry Non-equilibrium Flow Field

Using the air plasma ignition technique, physicochemical process of burning can be accelerated; concentration limit ofretrofires both can be extended; reliability of retrofires and stability of burning can be improved. In this paper, using internalequivalent heat area in place of electric are that created Ohm heat, the flow fields of thermodynamic equilibrium chemistry con-gealment and chemistry non-equilibrium in the plasma generator were simulated. The influences of the inlet prerotation angleof air, the inlet total pressure of air and the airflow compression angle of spray nozzle on the temperature on the surface of elec-

Turbulent Flow Action of Pulp in Wet - Laid Non - Woven Processes

The paper deals with the fluid field of web forming in wet-laid non-woven production.The influence of the turbulent flow on blending fiber and occluded fluid produced in pulp flow has been discussed in theory and practice.The suitable use of the imported velocity of pulp is very important in producing wet-laid products of good quality.

CYFRA 21-1 as an early predictor of first line chemotherapy response in advanced non small cell lung cancer

Objective： In an era of ever evolving, promising new therapies for advanced non small cell lung cancer （NSCLC）, early predictors of response to therapy, are needed. We evaluated early variations in CYFRA 21-1 serum levels of patients with advanced NSCLC receiving first line chemotherapy and correlated the results with objective tumor response. Methods： 29 consecutive, previously untreated, patients of advanced non small cell lung cancer, with measurable disease on CT scan were evaluated. All patients were treated with conventional systemic chemotherapy, although the choice of chemotherapy was left to the discretion of the treating physicians. Serum samples were obtained immediately before the start of 1st and 2nd cycles of chemotherapy. CYFRA 21-1 was measured with an electrochemiluminescense immunoassay on an automatic analyzer （Elecsys 2000; Roche Diagnostics）. Response was evaluated using Response evaluation criteria in solid tumors （RECIST） criteria. Results： 10 patients had partial response, 9 patients had stable disease and 9 had progressive disease. None of the patients had complete response. 21/29 （72%） patients had an elevated baseline value of CYFRA 21-1.62% patients （18/29） had a decrease in CYFRA 21-1 after 1 cycle of chemotherapy. The average reduction in the 2nd reading was irrespective of whether baseline value was normal or not. The average reduction was statistically significant （P = 0.002; 95% CI, from 0.8369 to 3.49464; t test）. 8 out of 10 （80%） patients with partial response had a reduction in their 2nd reading of. CYFRA （P = 0.019; 95% CI, from 0.81965 to 7.20035; t test） which was significant. We also observed that 6/9 （66%） patients whose disease remains stable also had a decrease in their subsequent reading （P = 0.0106; 95% CI, from -0.44942 to 3.82720; t test）, though it was not significant statistically. Although 5 out of 9 （55%） patients, who had an increase in their CYFRA 21-1 level, had progressive disease, but it was not statistically significant （P = 0.537; 95% CI, from -1.20021 to 2.13354; ttest）. 14 out of 19 （73%） who either had partial response or had stable disease, had a reduction in their 2nd value of CYFRA 21-1 and was significant statistically （P = 0.004; 95% CI, from 0.74792 to 3.50208; t test）. We also observed that except for 1 patient, all patients who had a decrease of 42% or more in their subsequent CYFRA 21-1 level, were those who had either responded to chemotherapy or had stable disease （P = 0.001）, which was statistically significant. Conclusion： We can conclude that monitoring of serum marker CYFRA 21-1, early dudng first-line chemotherapy may be a useful prognostic tool for evaluation of early tumor response in patients with advanced NSCLC.

Fused behavior recognition model based on attention mechanism

With the rapid development of deep learning technology,behavior recognition based on video streams has made great progress in recent years.However,there are also some problems that must be solved:(1)In order to improve behavior recognition performance,the models have tended to become deeper,wider,and more complex.However,some new problems have been introduced also,such as that their real-time performance decreases;(2)Some actions in existing datasets are so similar that they are difficult to distinguish.To solve these problems,the ResNet34-3DRes18 model,which is a lightweight and efficient two-dimensional(2D)and three-dimensional(3D)fused model,is constructed in this study.The model used 2D convolutional neural network(2DCNN)to obtain the feature maps of input images and 3D convolutional neural network(3DCNN)to process the temporal relationships between frames,which made the model not only make use of 3DCNN’s advantages on video temporal modeling but reduced model complexity.Compared with state-of-the-art models,this method has shown excellent performance at a faster speed.Furthermore,to distinguish between similar motions in the datasets,an attention gate mechanism is added,and a Res34-SE-IM-Net attention recognition model is constructed.The Res34-SE-IM-Net achieved 71.85%,92.196%,and 36.5%top-1 accuracy(The predicting label obtained from model is the largest one in the output probability vector.If the label is the same as the target label of the motion,the classification is correct.)respectively on the test sets of the HMDB51,UCF101,and Something-Something v1 datasets.

MiR-16-5p plays an inhibitory role in human non-small cell lung cancer through Fermitin family member 2

Increasing evidence indicates that aberrant expressions of some microRNAs are associated with cancer progression.However,the roles and biological mechanisms of miRNA-16-5p in human non-small cell lung cancer(NSCLC)are not to be well studied.Here,we validated that the expression of miR-16-5p was decreased significantly in NSCLC samples and cell lines.The correlation between the clinicopathological features of NSCLC and the miR-16-5p expression showed that the expression of miR-16-5p in non-small cell lung cancer was linked with the advanced TNM stage,positive lymph node metastasis,with short overall survival(OS).Also,a negative correlation between miR-16-5p and Fermitin family member 2(FERMT2)was observed,implying there may be a potential link about their regulation.The hypothesis was further confirmed by in-silico analysis and dual-luciferase reporter assay.Moreover,we demonstrated that the transfections of miR-16-5p mimics could alter some biological characteristics of NSCLC cells remarkably accomplished by the expression variance of FERMT2 in vitro and in vivo assays.Summarily,this study demonstrated that miR-16-5p,as a tumor suppression factor in NSCLC by targeting FERMT2,could serve as one promising biomarker in the prediction for NSCLC patients.

ATMS Based Information Fusion Target Recognition Method

The non-monotonic problem exited in information fusion systems is solved. Through the introducing of non-monotonic reasoning method, which was realized with ATMS, into the information fusion system, it gains the ability to process insufficient information with flexibility and non-monotonic behavior. In the simulation test of our system, our system manifests its ability of dealing the insufficient and contradictory information, which partly solves the decision dilemma brought out by the insufficient information in battle situations. The information fusion target recognition system can process the information in battle situation fast and with flexibility.

基于CBAM-GRU的通信信号自动调制识别

本文研究了一种基于卷积注意力机制模块(CBAM)与门控循环单元网络(GRU)结合的CBAM-GRU分类模型,用于非合作通信系统中的自动调制识别技术。将信号预处理后的时域幅度值、相位值以及I/Q值合并,转换为输入采样值矩阵,进入网络进行信号分类识别。使用无线电数据集RadioML2016.10a进行仿真实验,并将CBAM-GRU模型与卷积神经网络(CNN)、长短期记忆网络(LSTM)、GRU、卷积长短时深度神经网络(CLDNN)进行比较。实验结果表明:CBAM-GRU模型的分类识别率达到92.79%,相较于对比模型分别提高了8.52%、1.84%、1.75%、8.61%,比传统的CNN或LSTM模型,在处理信号时能够更有效地捕捉时空特征,从而提高识别精度。

面向智慧海洋场景的5G-A关键技术的研究与应用

智慧海洋是实现海洋强国战略的必由之路。5G-A技术是基于5G技术的增强和演进,其引入的基于NTN非地面网络的天地一体、通信感知一体、内生智能等全新的革命性技术,加快推动智慧海洋的发展。面向海域场景,通过对5G-A NTN网络天地一体关键技术的网络架构、时频同步、时序管理、传输调度和移动性管理进行分析,实现面向智慧海洋的5G-A海天一体组网方案以及5G专网+双星深度融合,实现5G+智慧海洋的信息通信全面升级;通过对5G-A通感一体关键技术的工作模式、通感架构、空口设计、目标特征智能识别进行分析,基于海域不同的场景特点和感知指标需求,分场景设计通感一体部署方案,实现海上船只智能精准识别,实时智能海事监控,有效保障海上航线安全。

基于RT-DETR改进的皮带运输机异物识别方法

凭借兼顾检测精度与速度的特点,YOLO近年来已成为煤炭等工业领域目标检测模型的佼佼者。然而,YOLO的检测性能受到置信度阈值和非极大值抑制阈值等超参数设定的影响。因此,本研究提出了一种基于改进的RT-D ETR带式输送机非煤异物检测模型。该模型无需置信度过滤和非极大值抑制,从而提升了检测精度。此外,针对RT-DETR参数量较大、难以在计算资源有限的边缘设备上部署的问题,我们设计了一种EMA-Faster Net骨干网络,并将颈部网络的AIFI模块替换为LPE-AIFI模块。最后,我们采用TensorRT进行加速,并将模型部署到Jetson Orin Nano边缘计算设备上。实验结果表明,改进后的RT-DETR模型与具有相似参数量的YOLOv8s相比,其召回率高出5.6%,平均类别精度高出4.3%;经TensorRT加速后,模型帧率可达26.4 FPS,满足了实时监测的要求。

基于灰度-梯度共生矩阵的串联故障电弧特征

为了获取非线性负载回路中串联故障电弧的特征,利用自行研制的故障电弧实验系统开展了变频器和工控机负载条件下的串联故障电弧实验。提出一种基于灰度-梯度共生矩阵的特征提取方法。采用前向差分法对电流信号进行预处理,并利用小波包技术对差分信号进行分解、重构,将重构信号按频率高低依次放入二维数组;二维数组中同时刻各频段的能量归一化后,将其转换为0~255的灰度值图像,对灰度值图像进行Wiener滤波并采用Laplace算子进行锐化和加强处理,取频率高于1 562.5 Hz的图像求解灰度-梯度共生矩阵。通过计算共生矩阵的15种特征量,筛选得到工控机和变频器负载条件下串联故障电弧的特征向量。将获取的故障电弧特征向量输入到支持向量机进行故障电弧识别测试,验证了上述故障电弧特征提取方法的有效性。

超声波辅助下合成香豆素-3-羧酸及其对有色金属的离子识别

以水杨醛为起始原料,在超声波辅助下,简便地合成了香豆素-3-羧酸(4)。利用紫外可见光光谱法和荧光光谱法研究了化合物(4)对一些有色金属离子的识别作用。结果表明,随着有色金属离子的加入,化合物(4)的紫外可见光光谱发生了红移和增色效应,荧光光谱发生猝灭现象;化合物(4)对Yb2+的识别作用最优,络合常数为7.5×106M-1。

1-萘胺二乙酸的无保护性介质流体室温燐光特性及金属离子的影响

含 1 萘胺二乙酸 (NADA)的水溶液仅以Na2 SO3作除氧剂 ,以TlNO3作重原子微扰剂 ,即能产生强而稳定的室温光 (RTP)信号 ,其最大激发和发射波长λex/λem 为 3 2 1/5 71nm ,浓度分别在 1.0× 10 - 6 ～ 1.0× 10 - 5mol/L和 2 .0× 10 - 7～ 1.0× 10 - 6 mol/L范围内与RTP强度呈良好线性关系 ,检测限为 1.2× 10 - 8mol/L。NADA作为一种n→π 跃迁型的发光性氨羧络合剂 ,当其与金属离子配位时 ,因其氮原子上的n电子参于配位 ,显著降低氮原子上的电荷密度 ,提高了分子内电荷转移 (ICT)态的能量 ,发光量子产率降低 ,从而荧 ( )光强度下降 。

模式识别对La_x Ba_(1-x)TiO_3纳米多晶制备工艺参数的设计

利用溶胶－凝胶法合成钡钛前驱体，经干燥、焙烧等工艺制备了一系列镧掺杂钛酸钡纳米多晶粉体，利用ＸＲＤ确定物相及原始晶粒尺寸；通过模式识别中非线性映照及逆映照方法以原始晶粒尺寸作为目标值将试样分为３类，并通过分类图对工艺参数进行设计，按设计结果进行实验验证，获得了预期结果的多晶粉体，说明模式识别的应用有利于克服传统“炒菜式”合成方法的盲目性．

基于优化k-mer频率的宏基因组聚类方法

k-mer频率是进行宏基因组分类时的一种重要的数字特征,然而该特征的维数随参数k的增加呈指数增长,利用该特征进行宏基因组分类易陷入"维数灾难"。为解决此问题,本文提出了一种基于优化k-mer频率的宏基因组DNA序列聚类方法。首先,提取DNA序列的kmer频率特征;其次,使用非负矩阵分解算法对DNA序列的k-mer频率特征进行优化;最后,利用模糊C均值算法进行聚类。将本文方法在包含有不同物种个数的模拟宏基因组数据上运行的结果表明,其能有效地克服现有宏基因组数据分类方法计算量大的缺点,且分类性能优于同类算法。