Automatic diagnosis of diabetic retinopathy using vision transformer based on wide-field optical coherence tomography angiography

Diabetic retinopathy(DR)is one of the major causes of visual impairment in adults with diabetes.Optical coherence tomography angiography(OCTA)is nowadays widely used as the golden criterion for diagnosing DR.Recently,wide-field OCTA(WF-OCTA)provided more abundant information including that of the peripheral retinal degenerative changes and it can contribute in accurately diagnosing DR.The need for an automatic DR diagnostic system based on WF-OCTA pictures attracts more and more attention due to the large diabetic population and the prevalence of retinopathy cases.In this study,automatic diagnosis of DR using vision transformer was performed using WF-OCTA images(12 mm×12 mm single-scan)centered on the fovea as the dataset.WF-OCTA images were automatically classified into four classes:No DR,mild nonproliferative diabetic retinopathy(NPDR),moderate to severe NPDR,and proliferative diabetic retinopathy(PDR).The proposed method for detecting DR on the test set achieves accuracy of 99.55%,sensitivity of 99.49%,and specificity of 99.57%.The accuracy of the method for DR staging reaches up to 99.20%,which has been proven to be higher than that attained by classical convolutional neural network models.Results show that the automatic diagnosis of DR based on vision transformer and WF-OCTA pictures is more effective for detecting and staging DR.

Research on PolSAR Image Classification Method Based on Vision Transformer Considering Local Information

In response to the problem of inadequate utilization of local information in PolSAR image classification using Vision Transformer in existing studies, this paper proposes a Vision Transformer method considering local information, LIViT. The method replaces image patch sequence with polarimetric feature sequence in the feature embedding, and uses convolution for mapping to preserve image spatial detail information. On the other hand, the addition of the wavelet transform branch enables the network to pay more attention to the shape and edge information of the feature target and improves the extraction of local edge information. The results in Wuhan, China and Flevoland, Netherlands show that considering local information when using Vision Transformer for PolSAR image classification effectively improves the image classification accuracy and shows better advantages in PolSAR image classification.

基于Vision Transformer的虹膜——人脸多特征融合识别研究

为了提高生物特征识别系统的准确性和鲁棒性,本文研究基于计算机视觉的虹膜—人脸多特征融合识别方法。本文对面部图像中虹膜区域进行提取以及预处理,采用对比度增强和归一化操作,加强了特征提取的一致性,提升了图像质量。为了获取丰富的深度特征,本文使用Vision Transformer模型对预处理后的虹膜和面部图像进行特征提取。利用多头注意力机制将虹膜和面部的多模态特征信息进行融合,再利用全连接层进行分类识别。试验结果表明,该方法识别性能优秀,识别准确性显著提升。

Dual-Path Vision Transformer用于急性缺血性脑卒中辅助诊断

急性缺血性脑卒中是由于脑组织血液供应障碍导致的脑功能障碍,数字减影脑血管造影(DSA)是诊断脑血管疾病的金标准。基于患者的正面和侧面DSA图像,对急性缺血性脑卒中的治疗效果进行分级评估,构建基于Vision Transformer的双路径图像分类智能模型DPVF。为了提高辅助诊断速度,基于EdgeViT的轻量化设计思想进行了模型的构建;为了使模型保持轻量化的同时具有较高的精度,提出空间-通道自注意力模块,促进Transformer模型捕获更全面的特征信息,提高模型的表达能力;此外,对于DPVF的两分支的特征融合,构建交叉注意力模块对两分支输出进行交叉融合,促使模型提取更丰富的特征,从而提高模型表现。实验结果显示DPVF在测试集上的准确率达98.5%,满足实际需求。

基于Vision Transformer的小麦病害图像识别算法

小麦白粉病、赤霉病和锈病是危害小麦产量的三大病害。为提高小麦病害图像的识别准确率,构建一种基于Vision Transformer的小麦病害图像识别算法。首先,通过田间拍摄的方式收集包含小麦白粉病、赤霉病和锈病3种病害在内的小麦病害图像,并对原始图像进行预处理,建立小麦病害图像识别数据集;然后,基于改进的Vision Transformer构建小麦病害图像识别算法,分析不同迁移学习方式和数据增强对模型识别效果的影响。试验可知,全参数迁移学习和数据增强能明显提高Vision Transformer模型的收敛速度和识别精度。最后,在相同时间条件下,对比Vision Transformer、AlexNet和VGG16算法在相同数据集上的表现。试验结果表明,Vision Transformer模型对3种小麦病害图像的平均识别准确率为96.81%,相较于AlexNet和VGG16模型识别准确率分别提高6.68%和4.94%。

基于MIDE个体寻优的采摘机器人轨迹规划仿真

为了提高移动式采摘机器人的作业效率,实现高效率高精度的采摘作业,提出改进DE算法下移动式采摘机器人轨迹规划方法。建立移动式采摘机器人的动力学模型,获取采摘机器人的动力学特性;根据该特性,建立基于双目视觉的针孔成像模型完成采摘机器人果实采摘经过的路径点收集;将得到的路径点作为种群个体,对DE算法实施改进,选取一种基于多种群移民的差分进化算法(MIDE)展开个体寻优并输出最优解,实现移动式采摘机器人的最优轨迹规划。实验结果表明,所提方法具有较高的收敛能力和搜索效率,能够有效提高机器人的工作效率和稳定性。

IDE基因表达水平在急性淋巴细胞白血病患儿中的预后意义

背景难治及复发的急性淋巴细胞白血病(ALL)患儿预后不良,寻找更多新的、具有预后价值的分子标志以协助判断预后具有重要意义。目的探讨ALL患儿中IDE基因表达水平及其与预后的相关性。设计回顾性队列研究。方法回顾性纳入首都医科大学附属北京儿童医院2020年1月1日至10月31日收治的ALL患儿作为测试队列,对患儿诊断时骨髓单个核细胞中的IDE基因表达水平行ROC曲线分析,并设定IDE界值,比较IDE高表达组和低表达组的预后差异;以河南省儿童医院2020年1月1日至10月31日收治的ALL患儿为验证队列,验证设定的IDE界值分组标准与预后的相关性。同时对临床生物学特征和早期治疗反应等行预后影响因素分析。主要结局指标无事件生存(EFS)。结果①本研究共纳入初治ALL患儿77例,测试队列37例,验证队列40例。②诊断时IDE基因表达水平与ALL预后相关,发生不良事件的患儿诊断时IDE基因表达水平显著高于无不良事件的患儿(P<0.001)。③ROC结果显示,IDE基因表达水平可预测ALL患儿的预后,AUC为0.961(P<0.001)。根据ROC曲线设定0.72为界值,测试队列中IDE高表达组(IDE≥0.72,n=19)的预后明显差于IDE低表达组(IDE<0.72,n=18),3年EFS分别为55.7%和100%(P=0.002)。在验证队列中,IDE高表达组(n=26)和低表达组(n=14)的3年EFS分别为64.3%和96.2%(P=0.009),差异有统计学意义。④多因素分析显示,临床危险度高危和IDE基因高表达是ALL患儿EFS独立预后因素,风险比分别为4.254(95%CI:1.080~17.554,P=0.039)和21.773(95%CI:2.632~180.125,P=0.004)。由这两个独立预后因素组成的复发风险指数的预测预后能力高于现行临床危险度分型(AUC分别为0.892和0.741,P=0.009)。结论IDE基因高表达与ALL患儿预后不良密切相关。

基于Vision Transformer与迁移学习的裤装廓形识别与分类

针对裤装廓形识别与分类模型的分类不准确问题,文章采用带有自注意力机制的Vision Transformer模型实现裤装廓形图像的分类,对于图片背景等无关信息对廓形识别的干扰,添加自注意力机制,增强有用特征通道。为防止因裤型样本数据集较少产生过拟合问题,可通过迁移学习方法对阔腿裤、喇叭裤、紧身裤、哈伦裤4种裤装廓形进行训练和验证,将改进的Vision Transformer模型与传统CNN模型进行对比实验,验证模型效果。实验结果表明:使用Vision Transformer模型在4种裤装廓形分类上的分类准确率达到97.72%,与ResNet-50和MobileNetV2模型相比均有提升,可为服装廓形的图像分类识别提供有力支撑,在实际服装领域中有较高的使用价值。

细粒度图像分类上Vision Transformer的发展综述

细粒度图像分类(fine-grained image classification,FGIC)一直是计算机视觉领域中的重要问题。与传统图像分类任务相比,FGIC的挑战在于类间对象极其相似,使任务难度进一步增加。随着深度学习的发展,Vision Transformer(ViT)模型在视觉领域掀起热潮,并被引入到FGIC任务中。介绍了FGIC任务所面临的挑战,分析了ViT模型及其特性。主要根据模型结构全面综述了基于ViT的FGIC算法,包括特征提取、特征关系构建、特征注意和特征增强四方面内容,对每种算法进行了总结,并分析了它们的优缺点。通过对不同ViT模型在相同公用数据集上进行模型性能比较,以验证它们在FGIC任务上的有效性。最后指出了目前研究的不足,并提出未来研究方向,以进一步探索ViT在FGIC中的潜力。

Collaborative positioning for swarms:A brief survey of vision,LiDAR and wireless sensors based methods

As positioning sensors,edge computation power,and communication technologies continue to develop,a moving agent can now sense its surroundings and communicate with other agents.By receiving spatial information from both its environment and other agents,an agent can use various methods and sensor types to localize itself.With its high flexibility and robustness,collaborative positioning has become a widely used method in both military and civilian applications.This paper introduces the basic fundamental concepts and applications of collaborative positioning,and reviews recent progress in the field based on camera,LiDAR(Light Detection and Ranging),wireless sensor,and their integration.The paper compares the current methods with respect to their sensor type,summarizes their main paradigms,and analyzes their evaluation experiments.Finally,the paper discusses the main challenges and open issues that require further research.

基于Vision Transformer的阿尔茨海默病分类研究

为了有效地提升对阿尔茨海默病(AD)的磁共振成像(MRI)图像分类准确率,提出一种LC(Layer-Cut)-ViT方法。该方法通过引入Vision Transformer(ViT)的自注意力机制对MRI图像进行层切分,使模型能更好地理解图像的全局信息,同时突出切片间的特征关系。此外,通过配准、颅骨分离算法提取MRI图像的脑部组织部分,进一步提升模型的性能。实验结果显示,所提方法对阿尔茨海默病的MRI图像具有较好的分类能力。

Frequency and associated factors of accommodation and non-strabismic binocular vision dysfunction among medical university students

AIM:To investigate the frequency and associated factors of accommodation and non-strabismic binocular vision dysfunction among medical university students.METHODS:Totally 158 student volunteers underwent routine vision examination in the optometry clinic of Guangxi Medical University.Their data were used to identify the different types of accommodation and nonstrabismic binocular vision dysfunction and to determine their frequency.Correlation analysis and logistic regression were used to examine the factors associated with these abnormalities.RESULTS:The results showed that 36.71%of the subjects had accommodation and non-strabismic binocular vision issues,with 8.86%being attributed to accommodation dysfunction and 27.85%to binocular abnormalities.Convergence insufficiency(CI)was the most common abnormality,accounting for 13.29%.Those with these abnormalities experienced higher levels of eyestrain(χ2=69.518,P<0.001).The linear correlations were observed between the difference of binocular spherical equivalent(SE)and the index of horizontal esotropia at a distance(r=0.231,P=0.004)and the asthenopia survey scale(ASS)score(r=0.346,P<0.001).Furthermore,the right eye's SE was inversely correlated with the convergence of positive and negative fusion images at close range(r=-0.321,P<0.001),the convergence of negative fusion images at close range(r=-0.294,P<0.001),the vergence facility(VF;r=-0.234,P=0.003),and the set of negative fusion images at far range(r=-0.237,P=0.003).Logistic regression analysis indicated that gender,age,and the difference in right and binocular SE did not influence the emergence of these abnormalities.CONCLUSION:Binocular vision abnormalities are more prevalent than accommodation dysfunction,with CI being the most frequent type.Greater binocular refractive disparity leads to more severe eyestrain symptoms.

基于Vision Transformer和迁移学习的家庭领域哭声识别

针对SVM等传统机器学习算法准确率低和当前使用CNN处理家庭领域哭声识别在不同婴儿间出现泛化能力差的问题,提出了一种基于Vision Transformer和迁移学习的婴儿哭声音频分类算法。首先,为实现数据集样本的扩增,采用了包括梅尔频谱转换和数据增强的数据预处理技术,进而达到了增强模型鲁棒性的目的。而后,在微调后的Vision Transformer模型上进行迁移学习训练,同时,训练过程中利用了LookAhead优化器来不断调整模型参数以避免过拟合,最终实验实现了对婴儿哭声音频的自动分类。实验结果表明,本实验模型相比其他深度学习模型具有更高的精确率和更快的收敛速度,同时还能有效地学习到婴儿哭声中更具区分性的特征。可以在新生儿监护、听力筛查和异常检测等领域中发挥重要作用。

Intelligent Risk-Identification Algorithm with Vision and 3D LiDAR Patterns at Damaged Buildings

Existingfirefighting robots are focused on simple storage orfire sup-pression outside buildings rather than detection or recognition.Utilizing a large number of robots using expensive equipment is challenging.This study aims to increase the efficiency of search and rescue operations and the safety offirefigh-ters by detecting and identifying the disaster site by recognizing collapsed areas,obstacles,and rescuers on-site.A fusion algorithm combining a camera and three-dimension light detection and ranging(3D LiDAR)is proposed to detect and loca-lize the interiors of disaster sites.The algorithm detects obstacles by analyzingfloor segmentation and edge patterns using a mask regional convolutional neural network(mask R-CNN)features model based on the visual data collected from a parallelly connected camera and 3D LiDAR.People as objects are detected using you only look once version 4(YOLOv4)in the image data to localize persons requiring rescue.The point cloud data based on 3D LiDAR cluster the objects using the density-based spatial clustering of applications with noise(DBSCAN)clustering algorithm and estimate the distance to the actual object using the center point of the clustering result.The proposed artificial intelligence(AI)algorithm was verified based on individual sensors using a sensor-mounted robot in an actual building to detectfloor surfaces,atypical obstacles,and persons requiring rescue.Accordingly,the fused AI algorithm was comparatively verified.

基于Vision Transformer和迁移学习的垃圾图像分类研究

为解决垃圾图像分类中分类准确率低及小样本类别性能差的问题,以生活垃圾图像为研究对象,以正确识别生活垃圾类别为研究目标,利用Vision Transformer模型为分类网络架构,使用迁移学习机制实现该模型在华为云垃圾分类数据集上的训练及分类推理。实验结果表明,基于注意力机制的分类模型相较于基于卷积结构的ResNet、DenseNet分类模型具有更高的分类准确率,可达96%,同时测试集的混淆矩阵表明Vision Transformer分类模型在样本不均衡数据集中对于小样本类别也具有较高的准确率,具有实际部署、推理的应用价值。

基于改进Vision Transformer的道岔故障智能诊断

道岔故障种类繁多,特征复杂,存在检测难、分类难等问题,导致故障排查效率低下,对铁路运输安全构成威胁。Vision Transformer模型在图像分类方面具有较高准确度,但是其处理的是图像块,而不是传统的像素级特征,在某些情况下可能会影响曲线局部信息的获取。针对上述情况,提出一种基于改进Vision Transformer模型的故障曲线分类算法。首先,对典型道岔故障及原因进行梳理分类,指出几种典型的道岔故障;其次,对使用道岔动作电流数据生成的图像尺寸进行调整并根据故障图像特点进行数据增强,使用ResNet网络取代原Vision Transformer模型中的故障图像分块机制进行特征提取,同时采用相对位置编码增强模型的适应性和泛化能力;最后,利用模型的多头自注意力机制,综合全局与局部信息进行分类,并得到分类权重。经过实验验证,本文道岔故障分类识别总体准确率达99.77%,各分类识别的平均精确率达99.78%,与原模型相比,在训练集和验证集上的识别精度分别提升了5.4%和2.4%。为了更好地理解模型的性能,采用Grad-CAM方法将迭代过程可视化,剖析了模型关注区域的变化过程,并在测试集上与VGG-16、DenseNet121等经典分类模型进行性能对比;通过ROC曲线评估分类效果,显示改进的模型取得更优结果。研究结果为道岔故障识别分类提供了新的理论支持,并为未来的研究提供了新的思路和方法。

基于改进Vision Transformer网络的农作物病害识别方法

基于DCNN模型的农作物病害识别方法在实验室环境下识别准确率高,但面对噪声时缺少鲁棒性.为了兼顾农作物病害识别的精度和鲁棒性,本文在标准ViT模型基础上加入增强分块序列化和掩码多头注意力,解决标准ViT模型缺乏局部归纳偏置和视觉特征序列的自注意力过于关注自身的问题.实验结果表明,本文的EPEMMSA-ViT模型对比标准ViT模型可以更高效的从零学习;当添加预训练权重训练网络时,EPEMMSA-ViT模型在数据增强的PlantVillage番茄子集上能够得到99.63%的分类准确率;在添加椒盐噪声的测试数据集上,对比ResNet50、DenseNet121、MobileNet和ConvNeXt的分类准确率分别提升了6.08%、9.78%、29.78%和12.41%;在添加均值模糊的测试数据集上,对比ResNet50、DenseNet121、MobileNet和ConvNeXt的分类准确率分别提升了18.92%、31.11%、20.37%和19.58%.

基于Vision Transformer-LSTM(ViTL)的多时序遥感影像农作物分类方法

针对当前遥感农作物分类研究中深度学习模型对光谱时间和空间信息特征采样不足,农作物提取仍然存在边界模糊、漏提、误提的问题,提出了一种名为视觉Transformer-长短期记忆递归神经网络(Vision Transformer-long short term memory,ViTL)的深度学习模型,ViTL模型集成了双路Vision-Transformer特征提取、时空特征融合和长短期记忆递归神经网络(LSTM)时序分类等3个关键模块,双路Vision-Transformer特征提取模块用于捕获图像的时空特征相关性,一路提取空间分类特征,一路提取时间变化特征;时空特征融合模块用于将多时特征信息进行交叉融合;LSTM时序分类模块捕捉多时序的依赖关系并进行输出分类。综合利用基于多时序卫星影像的遥感技术理论和方法,对黑龙江省齐齐哈尔市讷河市作物信息进行提取,研究结果表明,ViTL模型表现出色,其总体准确率(Overall Accuracy,OA)、平均交并比(Mean Intersection over Union,MIoU)和F1分数分别达到0.8676、0.6987和0.8175,与其他广泛使用的深度学习方法相比,包括三维卷积神经网络(3-D CNN)、二维卷积神经网络(2-D CNN)和长短期记忆递归神经网络(LSTM),ViTL模型的F1分数提高了9%~12%,显示出显著的优越性。ViTL模型克服了面对多时序遥感影像的农作物分类任务中的时间和空间信息特征采样不足问题,为准确、高效地农作物分类提供了新思路。

Vision-based fatigue crack detection using global motion compensation and video feature tracking

Fatigue cracks that develop in civil infrastructure such as steel bridges due to repetitive loads pose a major threat to structural integrity.Despite being the most common practice for fatigue crack detection,human visual inspection is known to be labor intensive,time-consuming,and prone to error.In this study,a computer vision-based fatigue crack detection approach using a short video recorded under live loads by a moving consumer-grade camera is presented.The method detects fatigue crack by tracking surface motion and identifies the differential motion pattern caused by opening and closing of the fatigue crack.However,the global motion introduced by a moving camera in the recorded video is typically far greater than the actual motion associated with fatigue crack opening/closing,leading to false detection results.To overcome the challenge,global motion compensation(GMC)techniques are introduced to compensate for camera-induced movement.In particular,hierarchical model-based motion estimation is adopted for 2D videos with simple geometry and a new method is developed by extending the bundled camera paths approach for 3D videos with complex geometry.The proposed methodology is validated using two laboratory test setups for both in-plane and out-of-plane fatigue cracks.The results confirm the importance of motion compensation for both 2D and 3D videos and demonstrate the effectiveness of the proposed GMC methods as well as the subsequent crack detection algorithm.

Artificial hawk-eye camera for foveated, tetrachromatic, and dynamic vision

With the rapid development of drones and autonomous vehicles, miniaturized and lightweight vision sensors that can track targets are of great interests. Limited by the flat structure, conventional image sensors apply a large number of lenses to achieve corresponding functions, increasing the overall volume and weight of the system.