A Hybrid Free-Form Deformation(HFFD)method is developed to improve shape preservation in mesh deformation for perforated surfaces,which traditional Free-Form Deformation(FFD)techniques struggle to handle effectively.T...A Hybrid Free-Form Deformation(HFFD)method is developed to improve shape preservation in mesh deformation for perforated surfaces,which traditional Free-Form Deformation(FFD)techniques struggle to handle effectively.The proposed method enables high-fidelity parameterized deformation for both flat and curved perforated surfaces while maintaining mesh quality with minimal geometric distortion.To evaluate its effectiveness,comparative studies between HFFD and conventional FFD methods are conducted,demonstrating superior performance in mesh quality and geometric fidelity.The HFFD-based framework is further applied to the Multidisciplinary Design Optimization(MDO)of a double-wall turbine blade leading edge.Results indicate an 11.6%increase in cooling efficiency and a 16.21%reduction in maximum stress.Additionally,compared to traditional geometry-based parameterization in MDO,the HFFD approach improves model processing efficiency by 84.15%and overall optimization efficiency by20.05%.These findings demonstrate HFFD's potential to significantly improve complex engineering design optimization by achieving precise shape preservation and improving computational efficiency.展开更多
Computing electrostatic interaction on non-cooperative targets with unknown meshes is crucial for electrostatic-based space on-orbit services.Although meshes for electrostatic interaction computations can be reconstru...Computing electrostatic interaction on non-cooperative targets with unknown meshes is crucial for electrostatic-based space on-orbit services.Although meshes for electrostatic interaction computations can be reconstructed from point clouds,they are usually too dense,leading to high computational costs.This paper presents an optimization method for converting dense meshes into optimal meshes,enabling fast and accurate computation of the electrostatic interaction by point clouds.First,the dense mesh reconstructed from point clouds is simplified into a coarse mesh using local operators.Second,the simplified mesh is refined by an iterative strategy that integrates a lightweight method of moments and an impedance matrix inheritance technique,ultimately yielding an optimal mesh for computing the electrostatic interaction.Simulation results show that our method effectively optimizes dense meshes,making electrostatic interaction computations using point clouds approximately 63.4 times more efficient than the previous method.展开更多
Recently, semantic segmentation has been widely applied toimage processing, scene understanding, and many others. Especially, indeep learning-based semantic segmentation, the U-Net with convolutionalencoder-decoder ar...Recently, semantic segmentation has been widely applied toimage processing, scene understanding, and many others. Especially, indeep learning-based semantic segmentation, the U-Net with convolutionalencoder-decoder architecture is a representative model which is proposed forimage segmentation in the biomedical field. It used max pooling operationfor reducing the size of image and making noise robust. However, instead ofreducing the complexity of the model, max pooling has the disadvantageof omitting some information about the image in reducing it. So, thispaper used two diagonal elements of down-sampling operation instead ofit. We think that the down-sampling feature maps have more informationintrinsically than max pooling feature maps because of keeping the Nyquisttheorem and extracting the latent information from them. In addition,this paper used two other diagonal elements for the skip connection. Indecoding, we used Subpixel Convolution rather than transposed convolutionto efficiently decode the encoded feature maps. Including all the ideas, thispaper proposed the new encoder-decoder model called Down-Sampling andSubpixel Convolution U-Net (DSSC-UNet). To prove the better performanceof the proposed model, this paper measured the performance of the UNetand DSSC-UNet on the Cityscapes. As a result, DSSC-UNet achieved89.6% Mean Intersection OverUnion (Mean-IoU) andU-Net achieved 85.6%Mean-IoU, confirming that DSSC-UNet achieved better performance.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
基金supported by the National Science and Technology Major Project,China(No.2017-II-0006-0019)the National Natural Science Foundation of China(No.52375266)the Shaanxi Science Foundation for Distinguished Young Scholars,China(No.2022JC-36)。
文摘A Hybrid Free-Form Deformation(HFFD)method is developed to improve shape preservation in mesh deformation for perforated surfaces,which traditional Free-Form Deformation(FFD)techniques struggle to handle effectively.The proposed method enables high-fidelity parameterized deformation for both flat and curved perforated surfaces while maintaining mesh quality with minimal geometric distortion.To evaluate its effectiveness,comparative studies between HFFD and conventional FFD methods are conducted,demonstrating superior performance in mesh quality and geometric fidelity.The HFFD-based framework is further applied to the Multidisciplinary Design Optimization(MDO)of a double-wall turbine blade leading edge.Results indicate an 11.6%increase in cooling efficiency and a 16.21%reduction in maximum stress.Additionally,compared to traditional geometry-based parameterization in MDO,the HFFD approach improves model processing efficiency by 84.15%and overall optimization efficiency by20.05%.These findings demonstrate HFFD's potential to significantly improve complex engineering design optimization by achieving precise shape preservation and improving computational efficiency.
基金supported by the National Natural Science Foundation of China(No.62003269).
文摘Computing electrostatic interaction on non-cooperative targets with unknown meshes is crucial for electrostatic-based space on-orbit services.Although meshes for electrostatic interaction computations can be reconstructed from point clouds,they are usually too dense,leading to high computational costs.This paper presents an optimization method for converting dense meshes into optimal meshes,enabling fast and accurate computation of the electrostatic interaction by point clouds.First,the dense mesh reconstructed from point clouds is simplified into a coarse mesh using local operators.Second,the simplified mesh is refined by an iterative strategy that integrates a lightweight method of moments and an impedance matrix inheritance technique,ultimately yielding an optimal mesh for computing the electrostatic interaction.Simulation results show that our method effectively optimizes dense meshes,making electrostatic interaction computations using point clouds approximately 63.4 times more efficient than the previous method.
文摘Recently, semantic segmentation has been widely applied toimage processing, scene understanding, and many others. Especially, indeep learning-based semantic segmentation, the U-Net with convolutionalencoder-decoder architecture is a representative model which is proposed forimage segmentation in the biomedical field. It used max pooling operationfor reducing the size of image and making noise robust. However, instead ofreducing the complexity of the model, max pooling has the disadvantageof omitting some information about the image in reducing it. So, thispaper used two diagonal elements of down-sampling operation instead ofit. We think that the down-sampling feature maps have more informationintrinsically than max pooling feature maps because of keeping the Nyquisttheorem and extracting the latent information from them. In addition,this paper used two other diagonal elements for the skip connection. Indecoding, we used Subpixel Convolution rather than transposed convolutionto efficiently decode the encoded feature maps. Including all the ideas, thispaper proposed the new encoder-decoder model called Down-Sampling andSubpixel Convolution U-Net (DSSC-UNet). To prove the better performanceof the proposed model, this paper measured the performance of the UNetand DSSC-UNet on the Cityscapes. As a result, DSSC-UNet achieved89.6% Mean Intersection OverUnion (Mean-IoU) andU-Net achieved 85.6%Mean-IoU, confirming that DSSC-UNet achieved better performance.
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。
