MSD-Net: Pneumonia Classification Model Based on Multi-Scale Directional Feature Enhancement

Computer-aided diagnosis of pneumonia based on deep learning is a research hotspot.However,there are some problems that the features of different sizes and different directions are not sufficient when extracting the features in lung X-ray images.A pneumonia classification model based on multi-scale directional feature enhancement MSD-Net is proposed in this paper.The main innovations are as follows:Firstly,the Multi-scale Residual Feature Extraction Module(MRFEM)is designed to effectively extract multi-scale features.The MRFEM uses dilated convolutions with different expansion rates to increase the receptive field and extract multi-scale features effectively.Secondly,the Multi-scale Directional Feature Perception Module(MDFPM)is designed,which uses a three-branch structure of different sizes convolution to transmit direction feature layer by layer,and focuses on the target region to enhance the feature information.Thirdly,the Axial Compression Former Module(ACFM)is designed to perform global calculations to enhance the perception ability of global features in different directions.To verify the effectiveness of the MSD-Net,comparative experiments and ablation experiments are carried out.In the COVID-19 RADIOGRAPHY DATABASE,the Accuracy,Recall,Precision,F1 Score,and Specificity of MSD-Net are 97.76%,95.57%,95.52%,95.52%,and 98.51%,respectively.In the chest X-ray dataset,the Accuracy,Recall,Precision,F1 Score and Specificity of MSD-Net are 97.78%,95.22%,96.49%,95.58%,and 98.11%,respectively.This model improves the accuracy of lung image recognition effectively and provides an important clinical reference to pneumonia Computer-Aided Diagnosis.

Importance of Features Selection,Attributes Selection,Challenges and Future Directions for Medical Imaging Data:A Review

In the area of pattern recognition and machine learning,features play a key role in prediction.The famous applications of features are medical imaging,image classification,and name a few more.With the exponential growth of information investments in medical data repositories and health service provision,medical institutions are collecting large volumes of data.These data repositories contain details information essential to support medical diagnostic decisions and also improve patient care quality.On the other hand,this growth also made it difficult to comprehend and utilize data for various purposes.The results of imaging data can become biased because of extraneous features present in larger datasets.Feature selection gives a chance to decrease the number of components in such large datasets.Through selection techniques,ousting the unimportant features and selecting a subset of components that produces prevalent characterization precision.The correct decision to find a good attribute produces a precise grouping model,which enhances learning pace and forecast control.This paper presents a review of feature selection techniques and attributes selection measures for medical imaging.This review is meant to describe feature selection techniques in a medical domainwith their pros and cons and to signify its application in imaging data and data mining algorithms.The review reveals the shortcomings of the existing feature and attributes selection techniques to multi-sourced data.Moreover,this review provides the importance of feature selection for correct classification of medical infections.In the end,critical analysis and future directions are provided.

Exploring evolutionary features of directed weighted hazard network in the subway construction

A better understanding of previous accidents is an effective way to reduce the occurrence of similar accidents in the future. In this paper, a complex network approach is adopted to construct a directed weighted hazard network(DWHN) to analyze topological features and evolution of accidents in the subway construction. The nodes are hazards and accidents, the edges are multiple relationships of these nodes and the weight of edges are occurrence times of repetitive relationships. The results indicate that the DWHN possesses the property of small-world with small average path length and large clustering coefficient, indicating that hazards have better connectivity and will spread widely and quickly in the network. Moreover,the DWHN has the property of scale-free network for the cumulative degree distribution follows a power-law distribution.It makes DWHN more vulnerable to target attacks. Controlling key nodes with higher degree, strength and betweenness centrality will destroy the connectivity of DWHN and mitigate the spreading of accidents in the network. This study is helpful for discovering inner relationships and evolutionary features of hazards and accidents in the subway construction.

Well production optimization using streamline features-based objective function and Bayesian adaptive direct search algorithm

Well production optimization is a complex and time-consuming task in the oilfield development.The combination of reservoir numerical simulator with optimization algorithms is usually used to optimize well production.This method spends most of computing time in objective function evaluation by reservoir numerical simulator which limits its optimization efficiency.To improve optimization efficiency,a well production optimization method using streamline features-based objective function and Bayesian adaptive direct search optimization(BADS)algorithm is established.This new objective function,which represents the water flooding potential,is extracted from streamline features.It only needs to call the streamline simulator to run one time step,instead of calling the simulator to calculate the target value at the end of development,which greatly reduces the running time of the simulator.Then the well production optimization model is established and solved by the BADS algorithm.The feasibility of the new objective function and the efficiency of this optimization method are verified by three examples.Results demonstrate that the new objective function is positively correlated with the cumulative oil production.And the BADS algorithm is superior to other common algorithms in convergence speed,solution stability and optimization accuracy.Besides,this method can significantly accelerate the speed of well production optimization process compared with the objective function calculated by other conventional methods.It can provide a more effective basis for determining the optimal well production for actual oilfield development.

Enhance the Performance of Directional Feature-based Palmprint Recognition by Directional Response Stability Measurement

Palmprint recognition is an emerging biometrics technology that has attracted increasing attention in recent years. Many palmprint recognition methods have been proposed, including traditional methods and deep learning-based methods. Among the traditional methods, the methods based on directional features are mainstream because they have high recognition rates and are robust to illumination changes and small noises. However, to date, in these methods, the stability of the palmprint directional response has not been deeply studied. In this paper, we analyse the problem of directional response instability in palmprint recognition methods based on directional feature. We then propose a novel palmprint directional response stability measurement (DRSM) to judge the stability of the directional feature of each pixel. After filtering the palmprint image with the filter bank, we design DRSM according to the relationship between the maximum response value and other response values for each pixel. Using DRSM, we can judge those pixels with unstable directional response and use a specially designed encoding mode related to a specific method. We insert the DRSM mechanism into seven classical methods based on directional feature, and conduct many experiments on six public palmprint databases. The experimental results show that the DRSM mechanism can effectively improve the performance of these methods. In the field of palmprint recognition, this work is the first in-depth study on the stability of the palmprint directional response, so this paper has strong reference value for research on palmprint recognition methods based on directional features.

Direction navigability analysis of geomagnetic field based on Gabor filter

Direction navigability analysis is a supplement to the navigability analysis theory, in which extraction of the direction suitable-matching features(DSMFs) determines the evaluation performance. A method based on the Gabor filter is proposed to estimate the direction navigability of the geomagnetic field. First,the DSMFs are extracted based on the Gabor filter’s responses.Second, in the view of pattern recognition, the classification accuracy in fault diagnosis is introduced as the objective function of the hybrid particle swarm optimization(HPSO) algorithm to optimize the Gabor filter’s parameters. With its guidance, the DSMFs are extracted. Finally, a direction navigability analysis model is established with the support vector machine(SVM), and the performances of the models under different objective functions are discussed. Simulation results show the parameters of the Gabor filter have a significant influence on the DSMFs, which, in turn, affects the analysis results of direction navigability. Moreover, the risk of misclassification can be effectively reduced by using the analysis model with optimal Gabor filter parameters. The proposed method is not restricted in geomagnetic navigation, and it also can be used in other fields such as terrain matching and gravity navigation.

RESEARCH FOR CLUSTERING OF FEATURE MANUFACTURING-ORIENTED

The following questions are discussed: feature cluster, feature clusterconcept and the reasoning formula. The defect based on approach direction and feed direction areanalyzed. Feature tool axis direction concept and its definition method are submitted. The featurefor practical part is also clustered by tool axis direction.

A NEW APPROACH TO THE DESIGN OF A SYSTEM FOR FEATURE EXTRACTION BASED ON HUMAN-COMPUTER COLLABORATIVE TACTIC

We are involved in an embarrassing situation that the limited capa-bility of automated feature extraction in digital photogrammetric systems cannotsatisfy the increasing needs for rapid acquisition of semantic information for appli-cations.Facing this challenge,a new tactic,Human-Computer Collaborative（HCC）tactic,and a corresponding new method,Operator-Object Directed（OOD）method,are proposed for the design of a system for feature extractionfrom large scale aerial images.We hold that in almost all technical complex sys-tems,full automation will be neither technically feasible nor socially acceptable.The system should be designed to optimize through the cooperative operation withtwo agents in the system:the human and the computer.

结合主动光源和改进YOLOv5s模型的夜间柑橘检测方法

【目的】解决夜间环境下遮挡和较小柑橘难以准确识别的问题,实现采摘机器人全天候智能化作业。【方法】提出一种结合主动光源的夜间柑橘识别方法。首先,通过分析主动光源下颜色特征不同的夜间柑橘图像,选择最佳的光源色并进行图像采集。然后,提出一种夜间柑橘检测模型BI-YOLOv5s,该模型采用双向特征金字塔网络(Bi-FPN)进行多尺度交叉连接和加权特征融合,提高对遮挡和较小果实的识别能力;引入Coordinate attention(CA)注意力机制模块,进一步加强对目标位置信息的提取;采用融入Transformer结构的C3TR模块,在减少计算量的同时更好地提取全局信息。【结果】本文提出的BI-YOLOv5s模型在测试集上的精准率、召回率、平均准确率分别为93.4%、92.2%和97.1%,相比YOLOv5s模型分别提升了3.2、1.5和2.3个百分点。在所采用的光源色环境下,模型对夜间柑橘识别的正确率为95.3%,相比白光环境下提高了10.4个百分点。【结论】本文提出的方法对夜间环境下遮挡和小目标柑橘的识别具有较高的准确性,可为夜间果蔬智能化采摘的视觉精准识别提供技术支持。

移动机器人视觉里程计技术研究综述

随着移动机器人技术不断发展,里程计技术已经成为移动机器人实现环境感知的关键技术,其发展水平对提高机器人的自主化和智能化具有重要意义。首先,系统阐述了同步定位与地图构建(Simultaneous localization and mapping,SLAM)中激光SLAM和视觉SLAM的发展近况,阐述了经典SLAM框架及其数学描述,简要介绍了3类常见相机的相机模型及其视觉里程计的数学描述。其次,分别对传统视觉里程计和深度学习里程计的研究进展进行系统阐述。对比分析了近10年来各类里程计算法的优势与不足。另外,对比分析了7种常用数据集的性能。最后,从精度、鲁棒性、数据集、多模态等方面总结了里程计技术面临的问题,从提高算法实时性、鲁棒性等方面展望了视觉里程计的发展趋势为:更加智能化、小型化新型传感器的发展;与无监督学习融合;语义表达技术的提高;集群机器人协同技术的发展。

基于链接关系预测的弯曲密集型商品文本检测

针对商品包装文本检测任务中弯曲密集型文本导致的错检、漏检问题,提出了一种由2个子网络组成的基于链接关系预测的文本检测框架(text detection network based on relational prediction,RPTNet)。在文本组件检测网络中,下采样采用卷积神经网络和自注意力并行的双分支结构提取局部和全局特征,并加入空洞特征增强模块(DFM)减少深层特征图在降维过程中信息的丢失;上采样采用特征金字塔与多级注意力融合模块(MAFM)相结合的方式进行多级特征融合以增强文本特征间的潜在联系,通过文本检测器从上采样输出的特征图中检测文本组件;在链接关系预测网络中,采用基于图卷积网络的关系推理框架预测文本组件间的深层相似度,采用双向长短时记忆网络将文本组件聚合为文本实例。为验证RRNet的检测性能,构建了一个由商品包装图片组成的文本检测数据集(text detection dataset composed of commodity packaging,CPTD1500)。实验结果表明:RPTNet不仅在公开文本数据集CTW-1500和Total-Text上取得了优异的性能,而且在CPTD1500数据集上的召回率和F值分别达到了85.4%和87.5%,均优于当前主流算法。

融合人体骨架和姿势信息特征的轻量级人体动作识别方法

针对人体动作识别任务中特征值选取不当导致识别率低、使用多模态数据导致训练成本高等问题,提出一种轻量级人体动作识别方法。首先使用OpenPose、PoseNet提取出人体骨架信息,使用BWT69CL传感器提取姿势信息;其次对数据进行预处理、特征融合,对人体动作进行深度学习分类识别;最后,为验证此方法的有效性,在公开数据集WISDM、UCIHAR、HASC和自建的人体动作数据集上进行实验验证,并使用改进的目标引导注意力机制(target-guided attention,TGA)–长短期记忆(long short term memory,LSTM)网络输出最终的分类结果。实验结果表明,在自建数据集下融合姿势和骨架特征达到99.87%准确率,相比于只使用姿势信息特征,识别准确率提高了约5.31个百分点;相比于只使用人体骨架特征,识别准确率提高了约1.87个百分点;在识别时间上相比于只使用姿势信息,识别时间降低了约29.73 s;相比于只使用人体骨架数据,识别时间降低了约9 s。使用该方法能及时有效地反映人体的运动意图,有助于提高人体动作和行为的识别准确率和训练效率。

新时期我国博物馆发展特征及博物馆强国建设思考

博物馆是现代文明的标志和产物,是建设、传播中华民族现代文明的重要载体。新时期以来,我国博物馆在免费开放政策和社会需求带动下,围绕博物馆强国建设目标,规模不断增长,功能日益拓展,服务体系日益健全,供给更加丰富,呈现综合化、特色化、现代化趋势,社会作用显著增强,成为中国式现代化建设的重要组成部分。但在发展进程中,也面临开放压力增大,支撑保障不足,冷热分化不均,国际影响力有待提高等问题和挑战。需要从质量和结构上,夯实巩固发展基础,优化均衡治理体系,加大优质产品供给,为满足人民群众美好生活需要,建设富强民主文明和谐美丽的社会主义现代化强国做出更大贡献。

用于单图像超分辨率的全局特征高效融合网络

现有图像超分辨率网络中普遍存在对层间特征利用水平较低的现象,使得在图像重建过程中有细节特征丢失,最终处理结果纹理模糊、图像质量欠佳。为此提出一种用于图像超分辨率的全局特征高效融合网络模型。主体使用对称卷积神经网络实现浅层特征的逐级提取,并结合Transformer完成浅层与深层特征的融合利用。设计的对称自指导残差模块可以在浅层网络实现不同层间特征更具表达性的融合,同时提升网络的特征提取能力;特征互导融合模块可以增强网络对浅层特征与深层特征的融合能力,促进更多的特征信息参与到细图像重建过程。在Set5、Set14、BSD100和Urban100数据集上同近年来的经典网络(HR、CARN、IMDN、MADNet、LBNet)进行性能对比,实验结果表明:所提网络模型在峰值信噪比上有所提升,并在视觉直观对比中取得了较好的图像超分辨率效果,可改善超分辨率图像质量欠佳的问题。

结合注意力和多路径融合的实时肺结节检测算法

现有单阶段目标检测算法在肺结节检测中结节检出不敏感,卷积神经网络(CNN)在特征提取时多次上采样导致微小结节特征提取困难、检测效果差,并且现存肺结节检测算法模型复杂,不利于实际应用部署落地。针对上述问题,提出一种结合注意力机制和多路径融合的实时肺结节检测算法,并在此基础上改进上采样算法,提升肺部结节的检测精度和模型推理速度,且模型的权重小容易部署。首先,在特征提取的主干网络部分融合通道和空间的混合注意力机制;其次,改进采样算法,提高生成特征图的质量;最后在加强特征提取网络部分,在不同路径之间建立通道,实现深层和浅层特征的融合,将不同尺度的语义和位置信息融合。在LUNA16数据集的实验结果表明,相较于原始YOLOv5s算法,所提算法的精确率、敏感度和平均精度分别提升9.5、6.9和8.7个百分点,帧率达到131.6 frame/s,模型权重文件仅有14.2 MB,表明了所提算法可以实时检测肺结节,并且精度远高于YOLOv3和YOLOv8等现有单阶段检测算法。

基于特征模式识别的农田监控系统研究

针对农田监控系统中存在的识别准确度不高、数据处理效率低等问题,提出了一种基于特征模式识别的农田监控系统。对农田监控系统的功能进行了设计,包括系统整体架构设计、数据采集与传输模块设计、特征提取与模式识别算法设计,结合用户界面设计实验与结果分析,验证了所提出系统的有效性和性能优势。对农田监控系统在实际农业生产中的应用前景进行了展望,并提出了未来改进方向和研究趋势。

基于改进3D U-Net模型的肺结节分割方法研究

由于肺部CT图像的特征信息复杂度较高,经典3D U-Net网络在肺结节分割方面准确率较低,存在误分割等问题。基于此,提出一种基于改进3D U-Net的网络模型。通过将加入了密集块的3D U-Net网络和双向特征网络(Bi-FPN)融合,提高了模型分割精度。同时采用深度监督训练机制,进一步提高了网络性能。在公开数据集LUNA-16上对模型进行比较实验和评估,结果显示,改进后的3D U-Net网络,Dice相似系数较原模型提高4%,分割精度为93.9%,敏感度为94.3%,证明该模型在肺结节分割精度及准确率方面具有一定的应用价值。

基于Geomagic Design Direct的截面特征提取与逆向建模

当前逆向工程已成为一种消化、吸收先进技术,实现产品的创新设计和快速开发的重要技术手段。随着各种特征技术的不断发展和引进,基于特征的逆向建模方法逐渐成为逆向工程新的发展方向。根据已有的建模方法,提出一种在二维草图模式下截面特征轮廓线提取与参数化设计的逆向建模新思路。并以Geomagic Design Direct软件为平台,实现了二维草图模式下特征轮廓线的提取与参数化再设计的逆向建模方法,为产品的快速建模和创新设计提供了一种新的途径。

基于随机森林的地震灾害直接经济损失评估研究——以中国西部地区为例

为快速评估地震直接经济损失,针对我国西部地区,尝试采用随机森林机器学习回归算法,以1993—2017年震害数据为基础,结合各年份经济数据与抗震设计数据,经特征选择与参数优化后,进行模型的训练与测试。实验结果表明,在减少模型输入特征的情况下,优化后的随机森林模型可得到更优的评估结果。通过删除含有缺失特征样本的数据预处理方法,评估模型的决定系数R2达到0.86,优于中值补齐缺失特征数据预处理下的评估模型,更适用于地震直接经济损失的评估。实例验证表明该模型评估结果与实际经济损失有较好的一致性,可为抗震救灾提供决策支持。

结合力导向图分布算法的特征加权深度嵌入聚类

聚类分析作为无监督学习领域的一个重要研究方向,是许多数据驱动应用的核心.但是高维数据特有的高维距离趋同特性,使得高维空间样本近邻结构遭到破坏,从而使得大量基于距离(基于近邻)的聚类算法性能急剧下降.目前,大量研究者认为,高维数据往往包含大量与任务不相关特征及相互关联的特征,其真实特征维度往往要比原始特征维度低很多.在学习样本低维等价表示上,基于深度自编码器的深度嵌入学习尽可能地保留重构信息.然而,现有此类方法往往需要聚类损失引导聚类,这虽然提高了聚类性能,但聚类损失与重构损失间的内在矛盾,限制了聚类性能的进一步提高.基于力导向图分布算法的降维算法则是尽可能保留近邻结构信息的基础上学习样本低维表示,但是高维距离趋同的特性使得此类算法较难准确获取样本高维近邻结构信息.本文在深度自编码器与力导向图分布算法的基础上引入特征加权思想,使模型在具有强大的低维等价表示能力及根据数据近邻结构凸显簇结构能力的同时考虑特征对聚类任务的适合程度.5个数据集上与最新高维聚类算法的对比实验充分证明了本文算法的合理性与优越性.