Automated retinal layer segmentation on optical coherence tomography image by combination of structure interpolation and lateral mean filtering

Segmentation of layers in retinal images obtained by optical coherence tomography(OCT)has become an important clinical tool to diagnose ophthalmic diseases.However,due to the sus-ceptibility to speckle noise and shadow of blood vessels etc.,the layer segmentation technology based on a single image still fail to reach a satisfactory level.We propose a combination method of structure interpolation and lateral mean filtering(SI-LMF)to improve the signal-to-noise ratio based on one retinal image.Before performing one-dimensional lateral mean filtering to remove noise,structure interpolation was operated to eliminate thickness fluctuations.Then,we used boundary growth method to identify boundaries.Compared with existing segmentations,the method proposed in this paper requires less data and avoids the influence of microsaccade.The automatic segmentation method was verified on the spectral domain OCT volume images obtained from four normal objects,which successfully identified the boundaries of 10 physio-logical layers,consistent with the results based on the manual determination.

Automated retinal layer segmentation in optical coherence tomography images with intraretinal fluid

We propose a novel retinal layer segmentation method to accurately segment 10 retinal layers in optical coherence tomography(OCT)images with intraretinal fluid.The method used a fan filter to enhance the linear information pertaining to retinal boundaries in an OCT image by reducing the effect of vessel shadows and fluid regions.A random forest classifier was employed to predict the location of the boundaries.Two novel methods of boundary redirection(SR)and similarity correction(SC)were combined to carry out boundary tracking and thereby accurately locate retinal layer boundaries.Experiments were performed on healthy controls and subjects with diabetic macular edema(DME).The proposed method required an average of 415 s for healthy controls and of 482 s for subjects with DME and achieved high accuracy for both groups of subjects.The proposed method requires a shorter running time than previous methods and also provides high accuracy.Thus,the proposed method may be a better choice for small training datasets.

A Semi-Vectorial Morphological Segmentation Multi-Component Images of Coumarins on Thin Layer Combined with Laser for Better Separation

In this work, we propose an approach for the separation of coumarins from thin-layer morphological segmentation based on the acquisition of multicomponent images integrating different types of coumarins. The first step is to make a segmentation by region, by thresholding, by contour, etc. of each component of the digital image. Then, we proceeded to the calculations of parameters of the regions such as the color standard deviation, the color entropy, the average color of the pixels, the eccentricity from an algorithm on the matlab software. The mean color values atR = 91.20 in red, atB = 213.21 in blue showed the presence of samidin in the extract. The color entropy values HG = 5.25 in green and HB = 4.04 in blue also show the presence of visnadine in the leaves of Desmodium adscendens. These values are used to consolidate the database of separation and discrimination of the types of coumarins. The relevance of our coumarin separation or coumarin recognition method has been highlighted compared to other methods, such as the one based on the calculation of frontal ratios which cannot discriminate between two coumarins having the same frontal ratio. The robustness of our method is proven with respect to the separation and identification of some coumarins, in particular samidin and anglicine.

Seismic and vibration control of segmented isolation layer in underground structure-diaphragm wall system

In this paper,a seismic and vibration reduction measure of subway station is developed by setting a segmented isolation layer between the sidewall of structure and the diaphragm wall.The segmented isolation layer consists of a rigid layer and a flexible layer.The rigid layer is installed at the joint section between the structural sidewall and slab,and the flexible layer is installed at the remaining sections.A diaphragm wall-segmented isolation layer-subway station structure system is constructed.Seismic and vibration control performance of the diaphragm wall-segmented isolation layer-subway station structure system is evaluated by the detailed numerical analysis.Firstly,a three-dimensional nonlinear time-history analysis is carried out to study the seismic response of the station structure by considering the effect of different earthquake motions and stiffness of segmented isolation layer.Subsequently,the vibration response of site under training loading is also studied by considering the influence of different train velocities and stiffness of the segmented isolation layer.Numerical results demonstrate that the diaphragm wall-segmented isolation layer-subway station structure system can not only effectively reduce the lateral deformation of station structure,but also reduce the tensile damage of the roof slab.On the other hand,the developed reduction measure can also significantly reduce the vertical peak displacements of site under training loading.

Characteristics of Nappes and Segmentation of the Longmen Mountains Thrust Belt, Western Sichuan Basin, China

In order to reveal the nature of the segmentation of Longmen Mountains Thrust Belt caused by the three nappes (Jiaoziding, Jiudingshan, and Baoxing Nappe), several methods are applied in this paper, including field investigation, seismic explanation and balanced crossed section, etc. Results of research reveal that nappes in Longmen Mountains vary in geometry, kinematics, and dynamics. Jiaoziding Nappe has generally behaved in a ductile manner, whereas Jiudingshan Nappe has been rigid, and the rheology of Baoxing Nappe has been intermediate between that of the other two nappes. The development of nappes has resulted in tectonic segmentation of Longmen Mountains: the main structural style of the northern segment is thrust faulting, with Jiaoziding Nappe representing a giant syncline. Given its ductility, it absorbed lots of stress, with the least amount of tectonic shortening in the SE part of the nappe. In the middle segment, the deformation is controlled by the rigid Jiudingshan Nappe, whose frontal area records lots of tectonic shortening. Deformation in the southern segment is intermediate in character between that of the other two segments, characterized by horizontal zonation, as demonstrated by fault development, and vertical stratification, which indicates that fault development was controlled by lithology.

Self-similar segmentation and multifractality of post-stack seismic data

Layering detection is an important step in petroleum engineering.Time series of post-stack seismic data and wire-line log data belong to subsurface layering.They exhibit multifractal properties with complex patterns because of the heterogeneity and different genetic properties in the earth layers.In a multifractal configuration,any piece of a series has a distinct Hurst exponent that reflects its nature and can be used for zone detection.Time series are post-stack seismic traces and wire-line log data near the well-bores.Self-similar Autoregressive Exogenous(SAE)model is a modified method which can place self-similar post-stack seismic and wire-line log segments across layers with the same lithology.The results satisfy the capability of layering identification from seismic data by SAE model.

Retinal layer segmentation using gradient feature calculation in OCT

Retinal diseases pose significant challenges to global healthcare systems,necessitating accurate and e±cient diagnostic methods.Optical Coherence Tomography(OCT)has emerged as a valuable tool for diagnosing and monitoring retinal conditions due to its noncontact and noninvasive nature.This paper presents a novel retinal layering method based on OCT images,aimed at enhancing the accuracy of retinal lesion diagnosis.The method utilizes gradient analysis to effectively identify and segment retinal layers.By selecting a column of pixels as a segmentation line and utilizing gradient information from adjacent pixels,the method initiates and proceeds with the layering process.This approach addresses potential issues arising from partial layer overlapping,minimizing deviations in layer segmentation.Experimental results demonstrate the e±cacy of the proposed method in accurately segmenting eight retinal boundaries,with an average absolute position deviation of 1.75 pixels.By providing accurate segmentation of retinal layers,this approach contributes to the early detection and management of ocular conditions,ultimately improving patient outcomes and reducing the global burden of vision-related ailments.

Straw layer burial to alleviate salt stress in silty loam soils: Impacts of straw forms

Salt stress can be alleviated by straw layer burial in the soil, but little is known of the appropriate form of the straw layer for optimal regulation of soil water and salinity because of the uncontrollability of field tests. Here, the following four straw forms with compaction thickness of 5 cm buried 40–45 deep were studied: no straw layer(CK), segmented straw(SL, 5 cm in length), straw pellet(SK), and straw powder(SF). The three straw forms(SL, SK and SF) significantly delayed the infiltration of irrigation water down the column profile by 71.20–134.3 h relative to CK and the migration velocity of the wetting front under SF was the slowest. It took longer for the wetting front to transcend SK than SL but shorter for it to reach the bottom of soil column after water crossed the straw layer. Compared with CK, the average volumetric water content in the 0–40 cm soil layer increased by 6.45% under SL, 1.77% under SK and 5.39% under SF. The desalination rates at the 0–40 and 0–100 cm soil layers increased by 5.85 and 3.76% under SL, 6.64 and 1.47% under SK and 5.97 and 4.82% under SF. However, there was no significant difference among straw forms in the 0–40 cm soil layer. Furthermore, the salt leaching efficiency(SLE, g mm^–1 h^–1) above the 40 cm layer under SL was 0.0097, being significantly higher than that under SF(0.0071) by 37.23%. Salt storage under SL, SK and SF in the 40–45 cm layer accounted for 4.50, 16.92 and 7.43% of total storage in the 1-m column profile. Cumulative evaporation under SL and SF decreased significantly by 41.20 and 49.00%, with both treatments having the most significant inhibition of salt accumulation(resalinization rate being 36.06 and 47.15% lower than CK) in the 0–40 cm soil layer. In conclusion, the different forms of straw layers have desalting effects under high irrigation level(446 mm). In particular, SL and SF performed better than SK in promoting deep salt leaching and inhibiting salt accumulation on the soil surface. However, SL was simpler to implement and its SLE was higher. Therefore, the segmented 5 cm straw can be recommended as an optimum physical form for establishing a straw layer for managing saline soils for crop production.

3D automatic segmentation method for retinal optical coherence tomography volume data using boundary surface enhancement

With the introduction of spectral-domain optical coherence tomography(SD-OCT),much larger image datasets are routinely acquired compared to what was possible using the previous generation of time-domain OCT.Thus,there is a critical need for the development of three-dimensional(3D)segmentation methods for processing these data.We present here a novel 3D automatic segmentation method for retinal OCT volume data.Brie°y,to segment a boundary surface,two OCT volume datasets are obtained by using a 3D smoothingfilter and a 3D differentialfilter.Their linear combination is then calculated to generate new volume data with an enhanced boundary surface,where pixel intensity,boundary position information,and intensity changes on both sides of the boundary surface are used simultaneously.Next,preliminary discrete boundary points are detected from the A-Scans of the volume data.Finally,surface smoothness constraints and a dynamic threshold are applied to obtain a smoothed boundary surface by correcting a small number of error points.Our method can extract retinal layer boundary surfaces sequentially with a decreasing search region of volume data.We performed automatic segmentation on eight human OCT volume datasets acquired from a commercial Spectralis OCT system,where each volume of datasets contains 97 OCT B-Scan images with a resolution of 496
512(each B-Scan comprising 512 A-Scans containing 496 pixels);experimental results show that this method can accurately segment seven layer boundary surfaces in normal as well as some abnormal eyes.

Video segmentation using Maximum Entropy Model

Detecting objects of interest from a video sequence is a fundamental and critical task in automated visual surveillance. Most current approaches only focus on discriminating moving objects by background subtraction whether or not the objects of interest can be moving or stationary. In this paper, we propose layers segmentation to detect both moving and stationary target objects from surveillance video. We extend the Maximum Entropy (ME) statistical model to segment layers with features, which are collected by constructing a codebook with a set of codewords for each pixel. We also indicate how the training models are used for the discrimination of target objects in surveillance video. Our experimental results are presented in terms of the success rate and the segmenting precision.

Apply GPCA to Motion Segmentation

In this paper, we present a motion segmentation approach based on the subspace segmentation technique, the genera-lized PCA. By incorporating the cues from the neighborhood of intensity edges of images, motion segmentation is solved under an algebra framework. Our main contribution is to propose a post-processing procedure, which can detect the boundaries of motion layers and further determine the layer ordering. Test results on real imagery have confirmed the validity of our method.

Bacillary layer detachment presenting with posterior scleritis:case report

Background:Bacillary layer detachment(BALAD)is a phenomenon characterized by fluid accumulation at the myoid region of the inner photoreceptor segments identifiable on optical coherence tomography(OCT)imaging.This finding has been recently described in patients with diverse primary diagnoses which share the common feature of serous exudation in the posterior pole.However,thus far there have been very few reports in the literature of BALAD in patients with posterior scleritis.Case Description:A 16-year-old male presented with unilateral vision changes that acutely worsened overnight to significant unilateral vision loss.He was eventually diagnosed with idiopathic posterior scleritis with associated BALAD on OCT.Similar to other reported cases of BALAD,he experienced anatomic restoration of the outer retina followed by good visual recovery after treatment with high dose steroid,ultimately with complete recovery of both retinal anatomy and vision within 4 months.Conclusions:This case provides further evidence that posterior scleritis can be a cause of BALAD.The rapid presentation and excellent visual and anatomical outcome of this case is entirely consistent with known descriptions of BALAD in a variety of other conditions,further supporting the categorization of BALAD as an entity which retinal specialists should be able to recognize as distinct from other forms of intraretinal fluid,retinal detachment,and retinoschisis.

分层逐段抽吸在急性高负荷血栓脑卒中取栓术中的应用价值

目的探讨分层逐段抽吸在急性高负荷血栓脑卒中取栓术中的应用价值。方法回顾性选择2020年4月至2023年6月在丽水市中心医院接受机械取栓治疗的前循环高负荷血栓大血管闭塞(LVO)性急性脑卒中患者90例,其中采用支架取栓39例,分层逐段抽吸51例。比较两组患者的手术相关资料、手术安全性及预后。结果分层逐段抽吸组患者血管成功再通率高于支架取栓组,穿刺至再通时间短于支架取栓组,7 d美国国立卫生研究院卒中量表评分低于支架取栓组,90 d预后良好率高于支架取栓组,差异均有统计学意义(均P<0.05)。两组患者血栓逃逸率、采取补救治疗率、术后症状性脑出血、90 d死亡率比较,差异均无统计学意义(均P>0.05)。结论对于前循环高负荷血栓LVO性急性脑卒中患者,分层逐段抽吸相较于支架取栓,手术时间更短、血管成功再通率及90 d预后良好率更高,有较好的临床应用价值。

基于端到端深度神经网络和图搜索的OCT图像视网膜层边界分割方法

视网膜层边界的形态变化是眼部视网膜疾病出现的重要标志,光学相干断层扫描(optical coherence tomography,OCT)图像可以捕捉其细微变化,基于OCT图像的视网膜层边界分割能够辅助相关疾病的临床判断.在OCT图像中,由于视网膜层边界的形态变化多样,其中与边界相关的关键信息如上下文信息和显著性边界信息等对层边界的判断和分割至关重要.然而已有分割方法缺乏对以上信息的考虑,导致边界不完整和不连续.针对以上问题,提出一种“由粗到细”的基于端到端深度神经网络和图搜索(graph search,GS)的OCT图像视网膜层边界分割方法,避免了非端到端方法中普遍存在的“断层”现象.在粗分割阶段,提出一种端到端的深度神经网络—注意力全局残差网络(attention global residual network,AGR-Net),以更充分和有效的方式提取上述关键信息.具体地,首先设计一个全局特征模块(global feature module,GFM),通过从图像的4个方向扫描以捕获OCT图像的全局上下文信息;其次,进一步将通道注意力模块(channel attention module,CAM)与全局特征模块串行组合并嵌入到主干网络中,以实现视网膜层及其边界的上下文信息的显著性建模,有效解决OCT图像中由于视网膜层形变和信息提取不充分所导致的误分割问题.在细分割阶段,采用图搜索算法去除AGR-Net粗分割结果中的孤立区域或和孔洞等,保持边界的固定拓扑结构和连续平滑,以实现整体分割结果的进一步优化,为医学临床的诊断提供更完整的参考.最后,在两个公开数据集上从不同的角度对所提出的方法进行性能评估,并与最新方法进行比较.对比实验结果也表明所提方法在分割精度和稳定性方面均优于现有方法.

改进D-Linknet的眼底视网膜血管分割

临床医生可通过观察眼底视网膜血管及其分支对人体是否患有疾病进行早期诊断,但由于视网膜中的血管错综复杂,模型在分割时会出现对微细血管分割精确度不足的问题。为此,提出一种结合残差模块Res2-net以及高效通道注意力机制(efficient channel attention,ECA)的D-Linknet模型。首先,利用Res2-net代替基础模型中的残差模块Res-net以提升每个网络层的感受野;其次,在Res2-net中添加一种结合压缩激励(squeeze and excitation,SE)和门通道(gated channel transformation,GCT)的注意力机制模块,改善处于复杂背景下的血管分割效果和效率;在网络的解码层加入ECA确保模型计算的性能,避免因降维导致的精度下降;最后,融合改进的模型输出图与掩膜图细化分割结果。在公开数据集DRIVE、STARE上进行分割实验,模型准确度(accuracy,AC)分别为97.11%、96.32%,灵敏度(sensitivity,SE)为84.55%、83.92%,曲线下方范围的面积(area under curve,AUC)为0.9873和0.9766,分割效果优于其他模型。实验证明了算法的可行性,为后续研究提供科学依据。

退役动力电池双层均衡方法研究

针对如何提高退役电池组均衡速度,提出了一种双层均衡电路,底层电路采用基于电感的Buck-Boost均衡电路结构,顶层电路采用可重构电路结构。根据开路电压(OCV)和电池荷电状态(SOC)曲线关系采用分段均衡控制策略。设计双层均衡电路的仿真实验,仿真实验结果表明:在电池参数相同情况下,与传统Buck-Boost均衡电路相比,双层电路均衡时间减少25%,有效提高了电池组均衡速度。在电池状态和均衡电路结构相同的情况下,与单一变量的均衡策略相比,SOC-电压分段均衡策略时间减少11%,验证了该均衡方案可行性。

多尺度特征金字塔融合的街景图像语义分割

针对街景图像语义分割任务中的目标尺寸差异大、多尺度特征难以高效提取的问题,本文提出了一种语义分割网络(LDPANet).首先,将空洞卷积与引入残差学习单元的深度可分离卷积结合,来优化编码器结构,在降低了计算复杂度的同时缓解梯度消失的问题.然后利用层传递的迭代空洞空间金字塔,将自顶向下的特征信息依次融合,提高了上下文信息的有效交互能力;在多尺度特征融合之后引入属性注意力模块,使网络抑制冗余信息,强化重要特征.再者,以通道扩展上采样代替双线插值上采样作为解码器,进一步提升了特征图的分辨率.最后,LDPANet方法在Cityscapes和CamVid数据集上的精度分别达到了91.8%和87.52%,与近几年网络模型相比,本文网络模型可以精确地提取像素的位置信息以及空间维度信息,提高了语义分割的准确率.

航电枢纽闸墩混凝土模块化施工实践

航电枢纽以航运为主,结合发电,挡水构筑物多为闸坝,施工期河道兼顾通航,一个枯期需完成构筑物的施工,施工工期紧。结合工程施工特点,通过闸墩混凝土分层分块设计,划分闸墩混凝土施工模块,将钢筋制作与安装,模板制作、安装与拆除,混凝土浇筑与养护等施工流程进行详细的拆分,形成各流程的关键控制点。模块化、标准化的施工有效提升了施工管理人员的管理能力,规范了工人的操作行为,降低了施工过程中的风险,保证了项目施工顺利、优质地完成。本文提出了航电枢纽工程闸墩混凝土模块化施工实施方案,使闸墩混凝土施工形成生产流水线,有效提升了航电枢纽工程闸墩混凝土施工精细化管理水平。

集成全尺度融合和循环注意力的医学图像分割网络

深度学习中的编解码网络在图像特征提取和分层特征融合方面具有卓越的性能,常被用于医学图像分割。但是,目前主流的编解码网络分割方法仍面临编码和解码阶段单一网络挖掘的图像特征信息不足,以及仅使用简单的跳跃连接而无法充分利用全尺度特征包含的粗粒度信息和细粒度信息等问题。为了解决上述问题,提出了一种集成全尺度融合和循环注意力的医学图像分割网络。首先,在U-Net编码器中加入了结合多层感知机(MLP)的卷积MLP模块来提取图像的全局特征信息,用于扩大编码器的特征感受野。其次,通过全尺度特征融合模块使得各尺度跳跃连接特征进行粗粒度信息和细粒度信息的有效融合,减小各尺度跳跃连接特征间的语义差异,突出图像的关键特征信息。最后,解码器通过提出的结合循环神经网络(RNN)和注意力机制的循环注意力解码模块(RADU)来逐级精细化图像特征信息,加强特征提取的同时避免信息冗余,并得到高精度分割结果。在4个数据集上将所提方法与主流较优的方法进行比较,所提方法在像素精度和骰子相似系数两个指标上的图像分割精度均有提高。因此,所提出的用于医学图像分割的编解码网络利用全尺度特征融合模块和循环注意力解码模块,能够获得较优异的高精度分割结果,并且模型具有良好的噪声鲁棒性和抗干扰能力。

多目标算法分层优化策略在开关磁阻电机中的应用

针对电机多参数多目标协同优化较为复杂的问题,提出了基于非支配排序遗传算法分层迭代优化的思想。首先,介绍定子分段混合励磁开关磁阻电机的设计流程和工作原理。其次,选择电机的待优化参数和优化目标,并引入Pearson相关系数分析电机参数与优化目标的相关性,根据相关性结果对待优化参数进行分层;建立各层优化参数与优化目标的非线性模型,将非线性目标模型引入多目标优化算法。最终,在Pareto前沿中选取最优个体,完成对电机结构参数和控制参数的分层迭代优化,确定电机的最优结构参数和控制参数,并通过有限元分析软件进行验证。相比较于初始模型,优化后电机的效率略有提高,平均转矩增加12.44%,转矩脉动减小64.96%。根据最优参数制造出实验样机,实验结果验证了优化设计的有效性和优越性。