Background:Traditional imaging approaches to keratoconus(KCN)have thus far failed to produce a standardized approach for diagnosis.While many diagnostic modalities and metrics exist,none have proven robust enough to b...Background:Traditional imaging approaches to keratoconus(KCN)have thus far failed to produce a standardized approach for diagnosis.While many diagnostic modalities and metrics exist,none have proven robust enough to be considered a gold standard.This study aims to introduce novel metrics to differentiate between KCN and healthy corneas using three-dimensional(3D)measurements of surface area and volume.Methods:This retrospective observational study examined KCN patients along with healthy control patients between the ages of 20 and 79 years old at the University of Maryland,Baltimore.The selected patients underwent a nine-line raster scan anterior segment optical coherence tomography(AS-OCT).ImageJ was used to determine the central 6 mm of each image and each corneal image was then divided into six 1 mm segments.Free-D software was then used to render the nine different images into a 3D model to calculate corneal surface area and volume.A two-tailed Mann-Whitney test was used to assess statistical significance when comparing these subsets.Results:Thirty-three eyes with KCN,along with 33 healthy control,were enrolled.There were statistically significant differences between the healthy and KCN groups in the metric of anterior corneal surface area(13.927 vs.13.991 mm^(2),P=0.046),posterior corneal surface area(14.045 vs.14.173 mm^(2),P<0.001),and volume(8.430 vs.7.773 mm3,P<0.001)within the central 6 mm.Conclusions:3D corneal models derived from AS-OCT can be used to measure anterior corneal surface area,posterior corneal surface area,and corneal volume.All three parameters are statistically different between corneas with KCN and healthy corneas.Further study and application of these parameters may yield new methodologies for the detection of KCN.展开更多
The three-dimensional（3D） visualization of the functional bundles in the peripheral nerve provides direct and detailed intraneural spatial information. It is useful for selecting suitable surgical methods to repair ...The three-dimensional（3D） visualization of the functional bundles in the peripheral nerve provides direct and detailed intraneural spatial information. It is useful for selecting suitable surgical methods to repair nerve defects and in optimizing the construction of tissue-engineered nerve grafts. However, there remain major technical hurdles in obtaining, registering and interpreting 2D images, as well as in establishing 3D models. Moreover, the 3D models are plagued by poor accuracy and lack of detail and cannot completely reflect the stereoscopic microstructure inside the nerve. To explore and help resolve these key technical problems of 3D reconstruction, in the present study, we designed a novel method based on re-imaging techniques and computer image layer processing technology. A 20-cm ulnar nerve segment from the upper arm of a fresh adult cadaver was used for acetylcholinesterase（ACh E） staining. Then, 2D panoramic images were obtained before and after ACh E staining under the stereomicroscope. Using layer processing techniques in Photoshop, a space transformation method was used to fulfill automatic registration. The contours were outlined, and the 3D rendering of functional fascicular groups in the long-segment ulnar nerve was performed with Amira 4.1 software. The re-imaging technique based on layer processing in Photoshop produced an image that was detailed and accurate. The merging of images was accurate, and the whole procedure was simple and fast. The least square support vector machine was accurate, with an error rate of only 8.25%. The 3D reconstruction directly revealed changes in the fusion of different nerve functional fascicular groups. In conclusion. The technique is fast with satisfactory visual reconstruction.展开更多
In order to study the behavior and interconnection of network devices,graphs structures are used to formulate the properties in terms of mathematical models.Mesh network(meshnet)is a LAN topology in which devices are ...In order to study the behavior and interconnection of network devices,graphs structures are used to formulate the properties in terms of mathematical models.Mesh network(meshnet)is a LAN topology in which devices are connected either directly or through some intermediate devices.These terminating and intermediate devices are considered as vertices of graph whereas wired or wireless connections among these devices are shown as edges of graph.Topological indices are used to reflect structural property of graphs in form of one real number.This structural invariant has revolutionized the field of chemistry to identify molecular descriptors of chemical compounds.These indices are extensively used for establishing relationships between the structure of nanotubes and their physico-chemical properties.In this paper a representation of sodium chloride(NaCl)is studied,because structure of NaCl is same as the Cartesian product of three paths of length exactly like a mesh network.In this way the general formula obtained in this paper can be used in chemistry as well as for any degree-based topological polynomials of three-dimensional mesh networks.展开更多
Mesh segmentation is one of the important issues in digital geometry processing. Region growing method has been proven to be a efficient method for 3D mesh segmentation. However, in mesh segmentation, feature line ext...Mesh segmentation is one of the important issues in digital geometry processing. Region growing method has been proven to be a efficient method for 3D mesh segmentation. However, in mesh segmentation, feature line extraction algorithm is computationally costly, and the over-segmentation problem still exists during region merging processing. In order to tackle these problems, a fast and efficient mesh segmentation method based on improved region growing is proposed in this paper. Firstly, the dihedral angle of each non-boundary edge is defined and computed simply, then the sharp edges are detected and feature lines are extracted. After region growing process is finished, an improved region merging method will be performed in two steps by considering some geometric criteria. The experiment results show the feature line extraction algorithm can obtain the same geometric information fast with less computational costs and the improved region merging method can solve over-segmentation well.展开更多
Morphological analyses are key outcome assessments for nerve regeneration studies but are historically limited to tissue sections.Novel optical tissue clearing techniques enabling three-dimensional imaging of entire o...Morphological analyses are key outcome assessments for nerve regeneration studies but are historically limited to tissue sections.Novel optical tissue clearing techniques enabling three-dimensional imaging of entire organs at a subcellular resolution have revolutionized morphological studies of the brain.To extend their applicability to experimental nerve repair studies we adapted these techniques to nerves and their motor and sensory targets in rats.The solvent-based protocols rendered harvested peripheral nerves and their target organs transparent within 24 hours while preserving tissue architecture and fluorescence.The optical clearing was compatible with conventional laboratory techniques,including retrograde labeling studies,and computational image segmentation,providing fast and precise cell quantitation.Further,optically cleared organs enabled three-dimensional morphometry at an unprecedented scale including dermatome-wide innervation studies,tracing of intramuscular nerve branches or mapping of neurovascular networks.Given their wide-ranging applicability,rapid processing times,and low costs,tissue clearing techniques are likely to be a key technology for next-generation nerve repair studies.All procedures were approved by the Hospital for Sick Children’s Laboratory Animal Services Committee(49871/9)on November 9,2019.展开更多
In the shape analysis community,decomposing a 3D shape intomeaningful parts has become a topic of interest.3D model segmentation is largely used in tasks such as shape deformation,shape partial matching,skeleton extra...In the shape analysis community,decomposing a 3D shape intomeaningful parts has become a topic of interest.3D model segmentation is largely used in tasks such as shape deformation,shape partial matching,skeleton extraction,shape correspondence,shape annotation and texture mapping.Numerous approaches have attempted to provide better segmentation solutions;however,the majority of the previous techniques used handcrafted features,which are usually focused on a particular attribute of 3Dobjects and so are difficult to generalize.In this paper,we propose a three-stage approach for using Multi-view recurrent neural network to automatically segment a 3D shape into visually meaningful sub-meshes.The first stage involves normalizing and scaling a 3D model to fit within the unit sphere and rendering the object into different views.Contrasting viewpoints,on the other hand,might not have been associated,and a 3D region could correlate into totally distinct outcomes depending on the viewpoint.To address this,we ran each view through(shared weights)CNN and Bolster block in order to create a probability boundary map.The Bolster block simulates the area relationships between different views,which helps to improve and refine the data.In stage two,the feature maps generated in the previous step are correlated using a Recurrent Neural network to obtain compatible fine detail responses for each view.Finally,a layer that is fully connected is used to return coherent edges,which are then back project to 3D objects to produce the final segmentation.Experiments on the Princeton Segmentation Benchmark dataset show that our proposed method is effective for mesh segmentation tasks.展开更多
In this paper, an elastic particle mesh （EPM） model is presented. It can be used like a cover to sketch images. EPM offers two advantages： first, when putting on a sketch image, it helps to repair disconnections on...In this paper, an elastic particle mesh （EPM） model is presented. It can be used like a cover to sketch images. EPM offers two advantages： first, when putting on a sketch image, it helps to repair disconnections on salient features. Second, it hides trivial details in the image, thus has the ability of decreasing over-segmentation when used with watershed transformation.展开更多
从三维Mesh数据中分割建筑物立面以识别对象,是三维场景理解的关键,但现有方法多依赖高成本的精细标注数据。针对该问题,提出了一种半监督学习方法,引入一种基于对比学习和一致性正则化的半监督语义分割(semi-supervised semantic segme...从三维Mesh数据中分割建筑物立面以识别对象,是三维场景理解的关键,但现有方法多依赖高成本的精细标注数据。针对该问题,提出了一种半监督学习方法,引入一种基于对比学习和一致性正则化的半监督语义分割(semi-supervised semantic segmentation based on contrastive learning and consistency regularization,SS_CC)方法,用于分割三维Mesh数据的建筑物立面。在SS_CC方法中,改进后的对比学习模块利用正负样本之间的类可分性,能够更有效地利用类特征信息;提出的基于特征空间的一致性正则化损失函数,从挖掘全局特征的角度增强了对所提取建筑物立面特征的鉴别力。实验结果表明,所提出的SS_CC方法在F1分数、mIoU指标上优于当前一些主流方法,且在建筑物的墙面和窗户上的分割效果相对更好。展开更多
Objective:The aims of this study were to investigate the clinical applicability of 3D segmentation in measuring cochlear anatomical parameters,explore factors that influence the insertion angle of cochlear implant ele...Objective:The aims of this study were to investigate the clinical applicability of 3D segmentation in measuring cochlear anatomical parameters,explore factors that influence the insertion angle of cochlear implant electrodes in patients with inner ear malformations,and determine the value of 3D segmentation in predicting cochlear implant electrode insertion depth by simulating electrode implantation in a reconstructed 3D model.Methods:Data from 208 temporal bone CT scans of patients with a variety of inner ear malformations(including the CH,IP-Ⅰ,IP-Ⅱ,and IP-Ⅲtypes)who underwent cochlear implantation at our center were retrospectively analyzed.Preoperative temporal bone CT data were subjected to three-dimensional(3D)segmentation of the cochlea with a 3D slicer.Results:Cochlear malformation types,including IP typesⅠ(42 ears),Ⅱ(278ears),Ⅲ(20 ears),and CH(65 ears),were diagnosed and measured in 208 preoperative CT datasets.Cochlear anatomical parameters and electrode length were correlated,which partially explained the variations in electrode insertion angle.The mean angle of implantation among the enrolled patients was 564.33°,and the mean implantation angle prediction error in the 3D segmentation was|23.74|°.Conclusion:Three-dimensional segmentation from temporal bone CT is valuable for surgeons,especially in treating patients with inner ear malformation.Such insights will help surgeons understand overall anatomical variations,predict electrode implantation depth,and complete preoperative imaging assessments for cochlear implant insertion depth in patients with inner ear malformations.展开更多
Background:Coronary artery disease(CAD)is a major global health concern requiring efficient and accurate diagnostic methods.Manual interpretation of coronary computed tomography angiography(CTA)images is time-consumin...Background:Coronary artery disease(CAD)is a major global health concern requiring efficient and accurate diagnostic methods.Manual interpretation of coronary computed tomography angiography(CTA)images is time-consuming and prone to interobserver variability,underscoring the need for automated segmentation and stenosis detection tools.Methods:This study presents a hybrid multi-scale 3D segmentation framework utilizing both 3D U-Net and Enhanced 3D U-Net architectures,designed to balance computational efficiency and anatomical precision.Processed CTA images from the ImageCAS dataset underwent data standardization,normalization,and augmentation.The framework applies ensemble learning to merge coarse and fine segmentation masks,followed by advanced post-processing techniques,including connected component analysis and centerline extraction,to refine vessel delineation.Stenosis regions are detected using the Enhanced 3D U-Net and morphological operations for accurate localization.Results:The proposed pipeline achieved near-perfect segmentation accuracy(0.9993)and a Dice similarity coefficient of 0.8539 for coronary artery delineation.Precision,recall,and F1 scores for stenosis detection were 0.8418,0.8289,and 0.8397,respectively.The dual-model approach demonstrated robust performance across varied anatomical structures and effectively localized stenotic regions,indicating clear superiority over conventional models.Conclusion:This hybrid framework enables highly reliable and automated coronary artery segmentation and stenosis detection from 3D CTA images.By reducing reliance on manual interpretation and enhancing diagnostic consistency,the proposed method holds strong potential to improve clinical workflows for CAD diagnosis and management.展开更多
The contact characteristics of the rough tooth surface during the meshing process are significantly affected by the lubrication state.The coupling effect of tooth surface roughness and lubrication on meshing character...The contact characteristics of the rough tooth surface during the meshing process are significantly affected by the lubrication state.The coupling effect of tooth surface roughness and lubrication on meshing characteristics of planetary gear is studied.An improved three-dimensional(3 D)anisotropic tooth surface roughness fractal model is proposed based on the experimental parameters.Considering asperity contact and elastohydrodynamic lubrication(EHL),the contact load and flexibility deformation of the tooth surface are derived,and the deformation compatibility equation of the 3 D loaded tooth contact analysis(3 D-LTCA)method is improved.The asperity of the tooth surface changes the system from EHL to mixed lubrication and reduces the stiffness of the oil film.Compared with the sun planet gear,the asperity has a greater effect on the meshing characteristics of the ring-planet gear.Compared with the proposed method,the comprehensive stiffness obtained by the traditional calculation method considering the lubrication effect is smaller,especially for the ring-planet gear.Compared with roughness,speed and viscosity,the meshing characteristics of planetary gears are most sensitive to torque.展开更多
Use of compressed mesh in parallel rendering architecture is still an unexplored area, the main challenge of which is to partition and sort the encoded mesh in compression-domain. This paper presents a mesh compressio...Use of compressed mesh in parallel rendering architecture is still an unexplored area, the main challenge of which is to partition and sort the encoded mesh in compression-domain. This paper presents a mesh compression scheme PRMC (Parallel Rendering based Mesh Compression) supplying encoded meshes that can be partitioned and sorted in parallel rendering system even in encoded-domain. First, we segment the mesh into submeshes and clip the submeshes’ boundary into Runs, and then piecewise compress the submeshes and Runs respectively. With the help of several auxiliary index tables, compressed submeshes and Runs can serve as rendering primitives in parallel rendering system. Based on PRMC, we design and implement a parallel rendering architecture. Compared with uncompressed representation, experimental results showed that PRMC meshes applied in cluster parallel rendering system can dramatically reduce the communication requirement.展开更多
BACKGROUND Split liver transplantation(SLT)is a complex procedure.The left-lateral and right tri-segment splits are the most common surgical approaches and are based on the Couinaud liver segmentation theory.Notably,t...BACKGROUND Split liver transplantation(SLT)is a complex procedure.The left-lateral and right tri-segment splits are the most common surgical approaches and are based on the Couinaud liver segmentation theory.Notably,the liver surface following right trisegment splits may exhibit different degrees of ischemic changes related to the destruction of the local portal vein blood flow topology.There is currently no consensus on preoperative evaluation and predictive strategy for hepatic segmental necrosis after SLT.AIM To investigate the application of the topological approach in liver segmentation based on 3D visualization technology in the surgical planning of SLT.METHODS Clinical data of 10 recipients and 5 donors who underwent SLT at Shenzhen Third People’s Hospital from January 2020 to January 2021 were retrospectively analyzed.Before surgery,all the donors were subjected to 3D modeling and evaluation.Based on the 3D-reconstructed models,the liver splitting procedure was simulated using the liver segmentation system described by Couinaud and a blood flow topology liver segmentation(BFTLS)method.In addition,the volume of the liver was also quantified.Statistical indexes mainly included the hepatic vasculature and expected volume of split grafts evaluated by 3D models,the actual liver volume,and the ischemia state of the hepatic segments during the actual surgery.RESULTS Among the 5 cases of split liver surgery,the liver was split into a left-lateral segment and right trisegment in 4 cases,while 1 case was split using the left and right half liver splitting.All operations were successfully implemented according to the preoperative plan.According to Couinaud liver segmentation system and BFTLS methods,the volume of the left lateral segment was 359.00±101.57 mL and 367.75±99.73 mL,respectively,while that measured during the actual surgery was 397.50±37.97 mL.The volume of segment IV(the portion of ischemic liver lobes)allocated to the right tri-segment was 136.31±86.10 mL,as determined using the topological approach to liver segmentation.However,during the actual surgical intervention,ischemia of the right tri-segment section was observed in 4 cases,including 1 case of necrosis and bile leakage,with an ischemic liver volume of 238.7 mL.CONCLUSION 3D visualization technology can guide the preoperative planning of SLT and improve accuracy during the intervention.The simulated operation based on 3D visualization of blood flow topology may be useful to predict the degree of ischemia in the liver segment and provide a reference for determining whether the ischemic liver tissue should be removed during the surgery.展开更多
The process of segmenting point cloud data into several homogeneous areas with points in the same region having the same attributes is known as 3D segmentation.Segmentation is challenging with point cloud data due to...The process of segmenting point cloud data into several homogeneous areas with points in the same region having the same attributes is known as 3D segmentation.Segmentation is challenging with point cloud data due to substantial redundancy,fluctuating sample density and lack of apparent organization.The research area has a wide range of robotics applications,including intelligent vehicles,autonomous mapping and navigation.A number of researchers have introduced various methodologies and algorithms.Deep learning has been successfully used to a spectrum of 2D vision domains as a prevailing A.I.methods.However,due to the specific problems of processing point clouds with deep neural networks,deep learning on point clouds is still in its initial stages.This study examines many strategies that have been presented to 3D instance and semantic segmentation and gives a complete assessment of current developments in deep learning-based 3D segmentation.In these approaches’benefits,draw backs,and design mechanisms are studied and addressed.This study evaluates the impact of various segmentation algorithms on competitiveness on various publicly accessible datasets,as well as the most often used pipelines,their advantages and limits,insightful findings and intriguing future research directions.展开更多
The numerical simulation of a three-dimensional semiconductor device is a fundamental problem in information science. The mathematical model is defined by an initialboundary nonlinear system of four partial differenti...The numerical simulation of a three-dimensional semiconductor device is a fundamental problem in information science. The mathematical model is defined by an initialboundary nonlinear system of four partial differential equations: an elliptic equation for electric potential, two convection-diffusion equations for electron concentration and hole concentration, and a heat conduction equation for temperature. The first equation is solved by the conservative block-centered method. The concentrations and temperature are computed by the block-centered upwind difference method on a changing mesh, where the block-centered method and upwind approximation are used to discretize the diffusion and convection, respectively. The computations on a changing mesh show very well the local special properties nearby the P-N junction. The upwind scheme is applied to approximate the convection, and numerical dispersion and nonphysical oscillation are avoided. The block-centered difference computes concentrations, temperature, and their adjoint vector functions simultaneously.The local conservation of mass, an important rule in the numerical simulation of a semiconductor device, is preserved during the computations. An optimal order convergence is obtained. Numerical examples are provided to show efficiency and application.展开更多
文摘Background:Traditional imaging approaches to keratoconus(KCN)have thus far failed to produce a standardized approach for diagnosis.While many diagnostic modalities and metrics exist,none have proven robust enough to be considered a gold standard.This study aims to introduce novel metrics to differentiate between KCN and healthy corneas using three-dimensional(3D)measurements of surface area and volume.Methods:This retrospective observational study examined KCN patients along with healthy control patients between the ages of 20 and 79 years old at the University of Maryland,Baltimore.The selected patients underwent a nine-line raster scan anterior segment optical coherence tomography(AS-OCT).ImageJ was used to determine the central 6 mm of each image and each corneal image was then divided into six 1 mm segments.Free-D software was then used to render the nine different images into a 3D model to calculate corneal surface area and volume.A two-tailed Mann-Whitney test was used to assess statistical significance when comparing these subsets.Results:Thirty-three eyes with KCN,along with 33 healthy control,were enrolled.There were statistically significant differences between the healthy and KCN groups in the metric of anterior corneal surface area(13.927 vs.13.991 mm^(2),P=0.046),posterior corneal surface area(14.045 vs.14.173 mm^(2),P<0.001),and volume(8.430 vs.7.773 mm3,P<0.001)within the central 6 mm.Conclusions:3D corneal models derived from AS-OCT can be used to measure anterior corneal surface area,posterior corneal surface area,and corneal volume.All three parameters are statistically different between corneas with KCN and healthy corneas.Further study and application of these parameters may yield new methodologies for the detection of KCN.
基金supported by the National Natural Science Foundation of China,No.30571913a grant from the Science and Technology Project of Guangdong Province of China,No.2013B010404019+1 种基金the Natural Science Foundation of Guangdong Province of China,No.9151008901000006the Medical Scientific Research Foundation of Guangdong Province of China,No.A2009173
文摘The three-dimensional（3D） visualization of the functional bundles in the peripheral nerve provides direct and detailed intraneural spatial information. It is useful for selecting suitable surgical methods to repair nerve defects and in optimizing the construction of tissue-engineered nerve grafts. However, there remain major technical hurdles in obtaining, registering and interpreting 2D images, as well as in establishing 3D models. Moreover, the 3D models are plagued by poor accuracy and lack of detail and cannot completely reflect the stereoscopic microstructure inside the nerve. To explore and help resolve these key technical problems of 3D reconstruction, in the present study, we designed a novel method based on re-imaging techniques and computer image layer processing technology. A 20-cm ulnar nerve segment from the upper arm of a fresh adult cadaver was used for acetylcholinesterase（ACh E） staining. Then, 2D panoramic images were obtained before and after ACh E staining under the stereomicroscope. Using layer processing techniques in Photoshop, a space transformation method was used to fulfill automatic registration. The contours were outlined, and the 3D rendering of functional fascicular groups in the long-segment ulnar nerve was performed with Amira 4.1 software. The re-imaging technique based on layer processing in Photoshop produced an image that was detailed and accurate. The merging of images was accurate, and the whole procedure was simple and fast. The least square support vector machine was accurate, with an error rate of only 8.25%. The 3D reconstruction directly revealed changes in the fusion of different nerve functional fascicular groups. In conclusion. The technique is fast with satisfactory visual reconstruction.
文摘In order to study the behavior and interconnection of network devices,graphs structures are used to formulate the properties in terms of mathematical models.Mesh network(meshnet)is a LAN topology in which devices are connected either directly or through some intermediate devices.These terminating and intermediate devices are considered as vertices of graph whereas wired or wireless connections among these devices are shown as edges of graph.Topological indices are used to reflect structural property of graphs in form of one real number.This structural invariant has revolutionized the field of chemistry to identify molecular descriptors of chemical compounds.These indices are extensively used for establishing relationships between the structure of nanotubes and their physico-chemical properties.In this paper a representation of sodium chloride(NaCl)is studied,because structure of NaCl is same as the Cartesian product of three paths of length exactly like a mesh network.In this way the general formula obtained in this paper can be used in chemistry as well as for any degree-based topological polynomials of three-dimensional mesh networks.
基金Supported by the National Natural Science Foundation of China(61272192,61379112)the NSFC-Guang dong Joint Fund(U1135003)
文摘Mesh segmentation is one of the important issues in digital geometry processing. Region growing method has been proven to be a efficient method for 3D mesh segmentation. However, in mesh segmentation, feature line extraction algorithm is computationally costly, and the over-segmentation problem still exists during region merging processing. In order to tackle these problems, a fast and efficient mesh segmentation method based on improved region growing is proposed in this paper. Firstly, the dihedral angle of each non-boundary edge is defined and computed simply, then the sharp edges are detected and feature lines are extracted. After region growing process is finished, an improved region merging method will be performed in two steps by considering some geometric criteria. The experiment results show the feature line extraction algorithm can obtain the same geometric information fast with less computational costs and the improved region merging method can solve over-segmentation well.
基金the German Research Foundation(DA 2255/1-1,to SCD).
文摘Morphological analyses are key outcome assessments for nerve regeneration studies but are historically limited to tissue sections.Novel optical tissue clearing techniques enabling three-dimensional imaging of entire organs at a subcellular resolution have revolutionized morphological studies of the brain.To extend their applicability to experimental nerve repair studies we adapted these techniques to nerves and their motor and sensory targets in rats.The solvent-based protocols rendered harvested peripheral nerves and their target organs transparent within 24 hours while preserving tissue architecture and fluorescence.The optical clearing was compatible with conventional laboratory techniques,including retrograde labeling studies,and computational image segmentation,providing fast and precise cell quantitation.Further,optically cleared organs enabled three-dimensional morphometry at an unprecedented scale including dermatome-wide innervation studies,tracing of intramuscular nerve branches or mapping of neurovascular networks.Given their wide-ranging applicability,rapid processing times,and low costs,tissue clearing techniques are likely to be a key technology for next-generation nerve repair studies.All procedures were approved by the Hospital for Sick Children’s Laboratory Animal Services Committee(49871/9)on November 9,2019.
基金supported by the National Natural Science Foundation of China (61671397).
文摘In the shape analysis community,decomposing a 3D shape intomeaningful parts has become a topic of interest.3D model segmentation is largely used in tasks such as shape deformation,shape partial matching,skeleton extraction,shape correspondence,shape annotation and texture mapping.Numerous approaches have attempted to provide better segmentation solutions;however,the majority of the previous techniques used handcrafted features,which are usually focused on a particular attribute of 3Dobjects and so are difficult to generalize.In this paper,we propose a three-stage approach for using Multi-view recurrent neural network to automatically segment a 3D shape into visually meaningful sub-meshes.The first stage involves normalizing and scaling a 3D model to fit within the unit sphere and rendering the object into different views.Contrasting viewpoints,on the other hand,might not have been associated,and a 3D region could correlate into totally distinct outcomes depending on the viewpoint.To address this,we ran each view through(shared weights)CNN and Bolster block in order to create a probability boundary map.The Bolster block simulates the area relationships between different views,which helps to improve and refine the data.In stage two,the feature maps generated in the previous step are correlated using a Recurrent Neural network to obtain compatible fine detail responses for each view.Finally,a layer that is fully connected is used to return coherent edges,which are then back project to 3D objects to produce the final segmentation.Experiments on the Princeton Segmentation Benchmark dataset show that our proposed method is effective for mesh segmentation tasks.
基金Project supported by National Natural Science Foundation of China（ Grant No. 60272081 ）
文摘In this paper, an elastic particle mesh （EPM） model is presented. It can be used like a cover to sketch images. EPM offers two advantages： first, when putting on a sketch image, it helps to repair disconnections on salient features. Second, it hides trivial details in the image, thus has the ability of decreasing over-segmentation when used with watershed transformation.
文摘从三维Mesh数据中分割建筑物立面以识别对象,是三维场景理解的关键,但现有方法多依赖高成本的精细标注数据。针对该问题,提出了一种半监督学习方法,引入一种基于对比学习和一致性正则化的半监督语义分割(semi-supervised semantic segmentation based on contrastive learning and consistency regularization,SS_CC)方法,用于分割三维Mesh数据的建筑物立面。在SS_CC方法中,改进后的对比学习模块利用正负样本之间的类可分性,能够更有效地利用类特征信息;提出的基于特征空间的一致性正则化损失函数,从挖掘全局特征的角度增强了对所提取建筑物立面特征的鉴别力。实验结果表明,所提出的SS_CC方法在F1分数、mIoU指标上优于当前一些主流方法,且在建筑物的墙面和窗户上的分割效果相对更好。
基金supported by the National Key Research and Development Program of China(grant no.2022YFC2402705)National Municipal Natural Science Foundation(grant no.82471161)Beijing Municipal Natural Science Foundation(grant no.7244308)。
文摘Objective:The aims of this study were to investigate the clinical applicability of 3D segmentation in measuring cochlear anatomical parameters,explore factors that influence the insertion angle of cochlear implant electrodes in patients with inner ear malformations,and determine the value of 3D segmentation in predicting cochlear implant electrode insertion depth by simulating electrode implantation in a reconstructed 3D model.Methods:Data from 208 temporal bone CT scans of patients with a variety of inner ear malformations(including the CH,IP-Ⅰ,IP-Ⅱ,and IP-Ⅲtypes)who underwent cochlear implantation at our center were retrospectively analyzed.Preoperative temporal bone CT data were subjected to three-dimensional(3D)segmentation of the cochlea with a 3D slicer.Results:Cochlear malformation types,including IP typesⅠ(42 ears),Ⅱ(278ears),Ⅲ(20 ears),and CH(65 ears),were diagnosed and measured in 208 preoperative CT datasets.Cochlear anatomical parameters and electrode length were correlated,which partially explained the variations in electrode insertion angle.The mean angle of implantation among the enrolled patients was 564.33°,and the mean implantation angle prediction error in the 3D segmentation was|23.74|°.Conclusion:Three-dimensional segmentation from temporal bone CT is valuable for surgeons,especially in treating patients with inner ear malformation.Such insights will help surgeons understand overall anatomical variations,predict electrode implantation depth,and complete preoperative imaging assessments for cochlear implant insertion depth in patients with inner ear malformations.
文摘Background:Coronary artery disease(CAD)is a major global health concern requiring efficient and accurate diagnostic methods.Manual interpretation of coronary computed tomography angiography(CTA)images is time-consuming and prone to interobserver variability,underscoring the need for automated segmentation and stenosis detection tools.Methods:This study presents a hybrid multi-scale 3D segmentation framework utilizing both 3D U-Net and Enhanced 3D U-Net architectures,designed to balance computational efficiency and anatomical precision.Processed CTA images from the ImageCAS dataset underwent data standardization,normalization,and augmentation.The framework applies ensemble learning to merge coarse and fine segmentation masks,followed by advanced post-processing techniques,including connected component analysis and centerline extraction,to refine vessel delineation.Stenosis regions are detected using the Enhanced 3D U-Net and morphological operations for accurate localization.Results:The proposed pipeline achieved near-perfect segmentation accuracy(0.9993)and a Dice similarity coefficient of 0.8539 for coronary artery delineation.Precision,recall,and F1 scores for stenosis detection were 0.8418,0.8289,and 0.8397,respectively.The dual-model approach demonstrated robust performance across varied anatomical structures and effectively localized stenotic regions,indicating clear superiority over conventional models.Conclusion:This hybrid framework enables highly reliable and automated coronary artery segmentation and stenosis detection from 3D CTA images.By reducing reliance on manual interpretation and enhancing diagnostic consistency,the proposed method holds strong potential to improve clinical workflows for CAD diagnosis and management.
基金Project(2024A1515240020)supported by the Guangdong Basic and Applied Basic Research Foundation,China。
文摘The contact characteristics of the rough tooth surface during the meshing process are significantly affected by the lubrication state.The coupling effect of tooth surface roughness and lubrication on meshing characteristics of planetary gear is studied.An improved three-dimensional(3 D)anisotropic tooth surface roughness fractal model is proposed based on the experimental parameters.Considering asperity contact and elastohydrodynamic lubrication(EHL),the contact load and flexibility deformation of the tooth surface are derived,and the deformation compatibility equation of the 3 D loaded tooth contact analysis(3 D-LTCA)method is improved.The asperity of the tooth surface changes the system from EHL to mixed lubrication and reduces the stiffness of the oil film.Compared with the sun planet gear,the asperity has a greater effect on the meshing characteristics of the ring-planet gear.Compared with the proposed method,the comprehensive stiffness obtained by the traditional calculation method considering the lubrication effect is smaller,especially for the ring-planet gear.Compared with roughness,speed and viscosity,the meshing characteristics of planetary gears are most sensitive to torque.
基金Project supported by the National Basic Research Program (973) of China (No. 2002CB312105), the National Natural Science Founda-tion of China (No. 60573074), the Natural Science Foundation of Shanxi Province, China (No. 20041040), Shanxi Foundation of Tackling Key Problem in Science and Technology (No. 051129), and Key NSFC Project of "Digital Olympic Museum" (No. 60533080), China
文摘Use of compressed mesh in parallel rendering architecture is still an unexplored area, the main challenge of which is to partition and sort the encoded mesh in compression-domain. This paper presents a mesh compression scheme PRMC (Parallel Rendering based Mesh Compression) supplying encoded meshes that can be partitioned and sorted in parallel rendering system even in encoded-domain. First, we segment the mesh into submeshes and clip the submeshes’ boundary into Runs, and then piecewise compress the submeshes and Runs respectively. With the help of several auxiliary index tables, compressed submeshes and Runs can serve as rendering primitives in parallel rendering system. Based on PRMC, we design and implement a parallel rendering architecture. Compared with uncompressed representation, experimental results showed that PRMC meshes applied in cluster parallel rendering system can dramatically reduce the communication requirement.
基金The Third People's Hospital of Shenzhen Scientific Research Project,No.G2021008 and No.G2022008Shenzhen Key Medical Discipline Construction Fund,No.SZXK079Shenzhen Science and Technology Research and Development Fund,No.JCYJ20190809165813331 and No.JCYJ20210324131809027。
文摘BACKGROUND Split liver transplantation(SLT)is a complex procedure.The left-lateral and right tri-segment splits are the most common surgical approaches and are based on the Couinaud liver segmentation theory.Notably,the liver surface following right trisegment splits may exhibit different degrees of ischemic changes related to the destruction of the local portal vein blood flow topology.There is currently no consensus on preoperative evaluation and predictive strategy for hepatic segmental necrosis after SLT.AIM To investigate the application of the topological approach in liver segmentation based on 3D visualization technology in the surgical planning of SLT.METHODS Clinical data of 10 recipients and 5 donors who underwent SLT at Shenzhen Third People’s Hospital from January 2020 to January 2021 were retrospectively analyzed.Before surgery,all the donors were subjected to 3D modeling and evaluation.Based on the 3D-reconstructed models,the liver splitting procedure was simulated using the liver segmentation system described by Couinaud and a blood flow topology liver segmentation(BFTLS)method.In addition,the volume of the liver was also quantified.Statistical indexes mainly included the hepatic vasculature and expected volume of split grafts evaluated by 3D models,the actual liver volume,and the ischemia state of the hepatic segments during the actual surgery.RESULTS Among the 5 cases of split liver surgery,the liver was split into a left-lateral segment and right trisegment in 4 cases,while 1 case was split using the left and right half liver splitting.All operations were successfully implemented according to the preoperative plan.According to Couinaud liver segmentation system and BFTLS methods,the volume of the left lateral segment was 359.00±101.57 mL and 367.75±99.73 mL,respectively,while that measured during the actual surgery was 397.50±37.97 mL.The volume of segment IV(the portion of ischemic liver lobes)allocated to the right tri-segment was 136.31±86.10 mL,as determined using the topological approach to liver segmentation.However,during the actual surgical intervention,ischemia of the right tri-segment section was observed in 4 cases,including 1 case of necrosis and bile leakage,with an ischemic liver volume of 238.7 mL.CONCLUSION 3D visualization technology can guide the preoperative planning of SLT and improve accuracy during the intervention.The simulated operation based on 3D visualization of blood flow topology may be useful to predict the degree of ischemia in the liver segment and provide a reference for determining whether the ischemic liver tissue should be removed during the surgery.
基金This research was supported by the BB21 plus funded by Busan Metropolitan City and Busan Institute for Talent and Lifelong Education(BIT)and a grant from Tongmyong University Innovated University Research Park(I-URP)funded by Busan Metropolitan City,Republic of Korea.
文摘The process of segmenting point cloud data into several homogeneous areas with points in the same region having the same attributes is known as 3D segmentation.Segmentation is challenging with point cloud data due to substantial redundancy,fluctuating sample density and lack of apparent organization.The research area has a wide range of robotics applications,including intelligent vehicles,autonomous mapping and navigation.A number of researchers have introduced various methodologies and algorithms.Deep learning has been successfully used to a spectrum of 2D vision domains as a prevailing A.I.methods.However,due to the specific problems of processing point clouds with deep neural networks,deep learning on point clouds is still in its initial stages.This study examines many strategies that have been presented to 3D instance and semantic segmentation and gives a complete assessment of current developments in deep learning-based 3D segmentation.In these approaches’benefits,draw backs,and design mechanisms are studied and addressed.This study evaluates the impact of various segmentation algorithms on competitiveness on various publicly accessible datasets,as well as the most often used pipelines,their advantages and limits,insightful findings and intriguing future research directions.
基金supported the Natural Science Foundation of Shandong Province(ZR2016AM08)Natural Science Foundation of Hunan Province(2018JJ2028)National Natural Science Foundation of China(11871312).
文摘The numerical simulation of a three-dimensional semiconductor device is a fundamental problem in information science. The mathematical model is defined by an initialboundary nonlinear system of four partial differential equations: an elliptic equation for electric potential, two convection-diffusion equations for electron concentration and hole concentration, and a heat conduction equation for temperature. The first equation is solved by the conservative block-centered method. The concentrations and temperature are computed by the block-centered upwind difference method on a changing mesh, where the block-centered method and upwind approximation are used to discretize the diffusion and convection, respectively. The computations on a changing mesh show very well the local special properties nearby the P-N junction. The upwind scheme is applied to approximate the convection, and numerical dispersion and nonphysical oscillation are avoided. The block-centered difference computes concentrations, temperature, and their adjoint vector functions simultaneously.The local conservation of mass, an important rule in the numerical simulation of a semiconductor device, is preserved during the computations. An optimal order convergence is obtained. Numerical examples are provided to show efficiency and application.
