基于视觉心理学的SFS流形算法研究

通过视觉方法获取形状在科学研究及工程实践中有着重要作用,灰度图像是最容易获取的视觉信息之一。目前由灰度重构的形状(SFS,shape from shading)受到了广泛的关注,但现有的算法稳定性差、效率低,一直阻碍着SFS方法的推广应用。受近期的视觉心理学研究成果的启发,针对SFS问题,从局部分析入手,即:在图像上,在每一灰度极值点附近,根据灰度单调变化的情况,形成一个椭圆形邻域,对应着表面凹凸区;在每一椭圆域上,形状可用抛物面逼近,其参数可由局部辐射度约束求解;整个图像被椭圆形邻域覆盖,这些覆盖可构成一个二维流形;借助于流形上单位分解函数,各局部抛物面粘合在一起,形成光滑的整体形状;总体形状在整幅图像的灰度约束下再进行全局优化,可得到最终的重构结果。给出了初步算例,算法的稳定性和效率均有大幅度提高。     

基于局部加密纯无网格法非线性Cahn-Hilliard方程的模拟

为数值求解描述不同物质间相位分离现象的高阶非线性Cahn-Hilliard(C-H)方程,发展了一种基于局部加密纯无网格有限点集法(local refinement finite pointset method,LR-FPM).其构造过程为:1)将C-H方程中四阶导数降阶为两个二阶导数,连续应用基于Taylor展开和加权最小二乘法的FPM离散空间导数;2)对区域进行局部加密和采用五次样条核函数以提高数值精度;3)局部线性方程组求解中准确施加含高阶导数Neumann边值条件.随后,运用LR-FPM求解有解析解的一维/二维C-H方程,分析粒子均匀分布/非均匀分布以及局部粒子加密情况的误差和收敛阶,展示了LR-FPM较网格类算法在非均匀布点情况下的优点.最后,采用LR-FPM对无解析解的一维/二维C-H方程进行了数值预测,并与有限差分结果相比较.数值结果表明,LR-FPM方法具有较高的数值精度和收敛阶,比有限差分法更易数值实现,能够准确展现不同类型材料间相位分离非线性扩散现象随时间的演化过程.     

基于改进格林元的压裂水平井动态分析方法

以建立高效的动态分析方法为出发点,以边单元作为求解点,改进传统的格林元方法,减少未知数和求解矩阵维度;并提出基于改进格林元的加密网格加密方法,保证考虑复杂裂缝网络的压裂水平井动态模拟的早期精度.退化模型与半解析解、数值模拟结果进行对比,验证本文基于加密网格的改进格林元方法的准确性和动态分析的高效性.最后进行动态响应的敏感性分析,结果表明：①格林元方法是一种高精度的动态模拟方法,将求解节点设置在网格的边上可以提高压裂水平井动态模拟的速度;②改进格林元方法的加密基于叠加原理,不需要通过插值近似,其求解精度高.在相同加密网格条件下,基于本文改进格林元方法的加密效果比有限差分加密效果更佳;③复杂裂缝导流能力、改造区渗透率提高倍数、改造区大小等参数对压裂水平井动态特征影响较大,在动态分析和参数反演时,应着重考虑这些因素的影响.     

多孔介质结构对渗流惯性效应的影响规律研究

本文通过局部展开算法研究了多孔介质中排列结构对非线性渗流规律的影响．计算结果表明：规则排列结构的渗流标度律和渗流方向相关,然而非规则排列结构的渗流标度律不依赖于渗流方向而只依赖于多孔介质的几何结构．伴随着排列结构的非规则程度增强,不同方向的渗流阻力变化规律的差异性减小．当流场输运区域较为复杂且不存在较宽直通道时,渗流阻力在较宽雷诺数范围内体现出较为明显的标度律特性．周期性强非规则排列结构中不同区域大小的模型却具有相似的渗流阻力的标度律特性,同时标度律特性与孔隙度的大小相关性不大．阻力非线性标度律由多孔介质结构的不同而取值为2至3之间．     

Learning Visible Thermal Person Re-Identification via Spatial Dependence and Dual-Constraint Loss

Visible thermal person re-identification (VT Re-ID) is the task of matching pedestrian images collected by thermal and visible light cameras. The two main challenges presented by VT Re-ID are the intra-class variation between pedestrian images and the cross-modality difference between visible and thermal images. Existing works have principally focused on local representation through cross-modality feature distribution, but ignore the internal connection of the local features of pedestrian body parts. Therefore, this paper proposes a dual-path attention network model to establish the spatial dependency relationship between the local features of the pedestrian feature map and to effectively enhance the feature extraction. Meanwhile, we propose cross-modality dual-constraint loss, which adds the center and boundary constraints for each class distribution in the embedding space to promote compactness within the class and enhance the separability between classes. Our experimental results show that our proposed approach has advantages over the state-of-the-art methods on the two public datasets SYSU-MM01 and RegDB. The result for the SYSU-MM01 is Rank-1/mAP 57.74%/54.35%, and the result for the RegDB is Rank-1/mAP 76.07%/69.43%.     

Numerical and Experimental Investigation of the Conjugate Heat Transfer for a High-Pressure Pneumatic Control Valve Assembly

This paper uses heat transfer experiments and computational fluid dynamics (CFD) simulations to investigate the conjugate heat transfer (CHT) in a high-pressure pneumatic control valve assembly. A heat transfer test rig was constructed, and time–temperature histories of five test points placed on the valve assembly’s outer surface were recorded for study validation. The Unsteady Reynolds-Averaged Navier–Stokes (URANS) CFD methods with the standard k-ε turbulence closure equations were adopted in the numerical computations. Polyhedral grids were used; time step and mesh convergence studies were conducted. Simulated and measured temperatures profile comparisons revealed a good agreement. The CHT results obtained from CFD showed huge velocity fields downstream of the valve throat and the vent hole. The airflow through the valve was icy, mainly in the supersonic flow areas. Low temperatures below 273.15 K were recorded on the internal and external walls of the valve assembly. The consistency of the measured data with the numerical results demonstrates the effectiveness of polyhedral grids in exploring the CHT using CFD methods. The local entropy production rate analysis revealed that irreversibility is mainly due to viscous dissipation. The current CHT investigation provides a potential basis for thermostress analysis and optimization.     

两个典型氧桥三核铁髥配合物[Fe_3O(O_2CCH_2OC_6H_5)_6(3H_2O)]和[Fe_3O(TIEO)_2(O_2CPh)_2Cl_3]的局域自旋和磁性质的理论探讨(英文)

用从头算波函数(UHF或UDFT波函数)代替ZILSH方法中的半经验波函数得到了ABLSH方法,接着用该方法研究了两个典型氧桥三核铁髥配合物[Fe3O(O2CCH2OC6H5)6(3H2O)]和[Fe3O(TIEO)2(O2CPh)2Cl3]的局域自旋和磁性质。通过计算得到的局域自旋结果和前人的具有可比性,同时所得的磁交换耦合常数和实验值很吻合。该方法可作为研究海森堡型磁性系统(HM)的新工具。     

Mesoporous materials for drug delivery

Research on mesoporous materials for biomedical purposes has experienced an outstanding increase during recent years. Since 2001, when MCM-41 was first proposed as drug-delivery system, silica-based materials, such as SBA-15 or MCM-48, and some metal-organic frameworks have been discussed as drug carriers and controlled-release systems. Mesoporous materials are intended for both systemic-delivery systems and implantable local-delivery devices. The latter application provides very promising possibilities in the field of bone-tissue repair because of the excellent behavior of these materials as bioceramics. This Minireview deals with the advances in this field by the control of the textural parameters, surface functionalization, and the synthesis of sophisticated stimuli-response systems.     

基于样本库的人脸图像处理方法

本文给出了一种基于样本库的人脸图像超分辨率方法,这一方法结合了人脸图像全局相似性和局部特征相似性的约束.局部结构的相似性将超分辨率解约束到人脸局部结构空间,不同尺度上的局部相似性指导搜索最优解,然后将局部结构重建的人脸图像投影到由人脸图像确定的特征人脸空间.实验结果表明该算法得到的人脸具有很好的准确性和视觉效果.     

Metal‐Organic Gels Based on Carboxyl‐Functionalized Benzimidazole and Their Stimuli Responsivenesses

Three multi‐responsiveness supramolecular metal‐organic gels (MOGs) have been prepared upon Ba(OAc)₂, CdSO₄·8H₂O and Pb(NO₃)₂ with a simple ligand (G17) based on a carboxyl‐functionalized benzimidazole derivative in alcoholic‐water solutions. The MOGs display the formation of well‐developed nanofibrillar networks composed of intertwined fibers which provide stability to gels structures through coordination, hydrogen bonding and π‐ π interactions characterized by using field emission scanning electron microscopy (FESEM), the fourier transform infrared (FT‐IR) spectroscopy and powder X‐ray diffraction (XRD) techniques. MOG‐1 shows good stimuli responsiveness toward the changes in K₂CrO₄, both MOG‐2 and MOG‐3 do good job toward the changes in Na₂S. Moreover, because these MOGs were formed easily by gelator with some heavy metal ion, such as Cd(II) and Pb(II), it might provide the basis for heavy metal ion capture and removal.