Search

The upcoming Geospace Dynamics Constellation (GDC) mission aims to investigate dynamic processes active in Earth's upper atmosphere and their local, regional, and global characteristics. Achieving this goal will involve resolving and distinguishing spatial and temporal variability of ionospheric and thermospheric (IT) structures in a quantitative manner. This, in turn, calls for the development of sophisticated algorithms that are optimal in combining information from multiple in-situ platforms. This manuscript introduces an implementation of the least-squares gradient calculation approach previously developed by J. De Keyser with the focus of its application to the GDC mission. This approach robustly calculates spatial and temporal gradients of IT parameters from in-situ measurements from multiple spacecraft that form a flexible constellation. The previous work by De Keyser, originally developed for analysis of Cluster data, focused on 3-D Cartesian geometry, while the current work extends the approach to spherical geometry suitable for missions in Low Earth Orbit (LEO). The algorithm automatically provides error bars for the estimated gradients as well as the scales over which the gradients are expected to be constant. We evaluate the performance of the software on outputs of high-resolution global ionospheric/thermospheric simulations. It is shown that the software will be a powerful tool to explore GDC's ability to answer science questions that require gradient calculations. The code can also be employed in support of Observing System Simulation Experiments to evaluate suitability of various constellation geometries and assess the impact of measurement sensitivities on addressing GDC's science objectives. [ABSTRACT FROM AUTHOR]

The upcoming Geospace Dynamics Constellation (GDC) mission aims to investigate dynamic processes active in Earth’s upper atmosphere and their local, regional, and global characteristics. Achieving this goal will involve resolving and distinguishing spatial and temporal variability of ionospheric and thermospheric (IT) structures in a quantitative manner. This, in turn, calls for the development of sophisticated algorithms that are optimal in combining information from multiple in-situ platforms. This manuscript introduces an implementation of the least-squares gradient calculation approach previously developed by J. De Keyser with the focus of its application to the GDC mission. This approach robustly calculates spatial and temporal gradients of IT parameters from in-situ measurements from multiple spacecraft that form a flexible constellation. The previous work by De Keyser, originally developed for analysis of Cluster data, focused on 3-D Cartesian geometry, while the current work extends the approach to spherical geometry suitable for missions in Low Earth Orbit (LEO). The algorithm automatically provides error bars for the estimated gradients as well as the scales over which the gradients are expected to be constant. We evaluate the performance of the software on outputs of high-resolution global ionospheric/thermospheric simulations. It is shown that the software will be a powerful tool to explore GDC’s ability to answer science questions that require gradient calculations. The code can also be employed in support of Observing System Simulation Experiments to evaluate suitability of various constellation geometries and assess the impact of measurement sensitivities on addressing GDC’s science objectives.

Improvements in structural biology methods, in particular crystallography and cryo-electron microscopy, have created an increased demand for the refinement of atomic models against low-resolution experimental data. One way to compensate for the lack of high-resolution experimental data is to use a priori information about model geometry that can be utilized in refinement in the form of stereochemical restraints or constraints. Here, the definition and calculation of the restraints that can be imposed on planar atomic groups, in particular the angle between such groups, are described. Detailed derivations of the restraint targets and their gradients are provided so that they can be readily implemented in other contexts. Practical implementations of the restraints, and of associated data structures, in the Computational Crystallography Toolbox (cctbx) are presented.

Copyright of Journal of National University of Defense Technology / Guofang Keji Daxue Xuebao is the property of NUDT Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Recently, deep learning has made great achievements in facial expression recognition. However, occlusion and large skew will greatly affect the accuracy of facial expression recognition in practice. Therefore, we propose a novel framework based on symmetric SURF and heterogeneous soft partition network to quickly recognize facial recognition under partial occlusion. In this framework, an occlusion detection module based on symmetric SURF is presented to detect the occlusion part, which helps to locate the horizontal symmetric area of the occlusion area. After that, a face inpainting module based on mirror transition is presented to rapidly accomplish the face inpainting under the unsupervised circumstance. Moreover, a recognition network based on heterogeneous soft partitioning is proposed for the facial expression recognition. After heterogeneous soft partitioning, the weights of each part are input and to into the recognition network as more prior information for training. Finally, we feed the weighted image into the trained neural network for expression recognition. Experimental results show that the accuracy of the proposed method is respectively 7% and 8% higher than the average accuracies from the state-of-the-art methods on Cohn-Kanade (CK +) and fer2013 datasets. Besides, the run time of our method is 2.38 s faster than the most advanced. [ABSTRACT FROM AUTHOR]

Due to the complicated procedures, lack of standardized procedures, error-pronemanual operation andvisual interpretation of traditional microscopic microbiological detection methods, it is difficult to trace the results of microscopic detection, and thus the detection efficiency is low, so the development of microbial detection has encountered a bottleneck. In this era of information technology, it is an urgent problem to be solved on how to combine image processing technology with microbial detection to manage microbial samples more efficiently. The Histogram of Oriented Gradient (HOG) feature is a feature descriptor used for object detection in computer vision and image processing, which constructs the feature by calculating and counting the histogram of gradient direction of local area. In this paper, a method of microbial feature extraction based on Hog feature extraction is proposed, which adopts the color space normalization, gradient calculation, histogram of oriented gradient, normalization and feature extraction of overlapping blockshistogram. The experiment verifies that such method in this paper can be better used to identify and extract the effective feature points of the microbe image under the microscope, which provides a good foundation for the subsequent further processing of relevant professionals. [ABSTRACT FROM AUTHOR]

A method of balancing the optimize accuracy and efficiency of the mesh to achieve mesh adaptive fast optimization is proposed to solve the problem of low efficiency of mesh optimization in the existing 3D reconstruction based on images. After calculating the initial value of the vertex gradient along the triangular normal vector based on the image gray information, the mesh of each triangle is marked differently according to the different results: The active triangles and the lazy triangles, through the active in the mesh is refined to abandon the optimization of the lazy triangles in exchanging for improvement of efficiency. The optimization of the efficiency is fast improving while obtaining better mesh vertices. We choose classical images church and fountain data in the 3D reconstruction experiment of computer vision and the widely used of unmanned aerial vehicle (UAV) images in photogrammetry to reconstruct mesh to refine. The mesh model optimization time and precision comparison results show that this algorithm can greatly improve the mesh optimize efficiency to obtain the 3D model faster. [ABSTRACT FROM AUTHOR]

Highlights • AUV that has strong coupling relationship among disciplines is optimized. • Three kinds of software are integrated to calculate drag automatically. • The analytic methods are introduced to calculate gradients efficiently. • Optimization is performed in a multi-start approach to find the global optimum. Abstract In this paper, multidisciplinary design optimization (MDO) is introduced for the conceptual design of an autonomous underwater vehicle (AUV). The purpose is to minimize the energy consumption with predefined sailing distance. AUV is decomposed into three disciplines and the coupling relationship is analyzed to build the optimization model. Since drag plays a major role in energy consumption, a hydrodynamic analysis framework is established for drag calculation, consisting of parametric modeling, mesh auto-generation and numerical simulation. In order to complete the optimization effectively, multidisciplinary feasible (MDF) architecture is used and gradient-based optimization algorithm is adopted. Moreover, analytic methods are incorporated into the MDF architecture via gradient to further improve the efficiency of gradient calculation. The optimization result shows that the optimized AUV is much more energy-saving than the initial design and the MDF architecture via coupled analytic methods is quite efficient. [ABSTRACT FROM AUTHOR]

Full waveform inversion (FWI) is a challenging procedure due to the high computational cost related to the modeling, especially for the elastic case. The graphics processing unit (GPU) has become a popular device for the high-performance computing (HPC). To reduce the long computation time, we design and implement the GPU-based 2D elastic FWI (EFWI) in time domain using a single GPU card. We parallelize the forward modeling and gradient calculations using the CUDA programming language. To overcome the limitation of relatively small global memory on GPU, the boundary saving strategy is exploited to reconstruct the forward wavefield. Moreover, the L-BFGS optimization method used in the inversion increases the convergence of the misfit function. A multiscale inversion strategy is performed in the workflow to obtain the accurate inversion results. In our tests, the GPU-based implementations using a single GPU device achieve >15 times speedup in forward modeling, and about 12 times speedup in gradient calculation, compared with the eight-core CPU implementations optimized by OpenMP. The test results from the GPU implementations are verified to have enough accuracy by comparing the results obtained from the CPU implementations. [ABSTRACT FROM AUTHOR]

The topology optimization of members of flexible multibody systems is considered for energy-efficient lightweight design, where the gradient calculation has an essential role. In topology optimization of flexible multibody systems, where the function evaluations are very time consuming, the gradient information is necessary to accelerate the optimization process. Different approaches have been introduced and tested for the gradient calculation in the fully coupled topology optimization of flexible multibody systems. However, the computation and implementation costs of these methods are high, which limits the optimization size and the possible number of design variables. In this work, we present a modified gradient calculation based on the equivalent static load (ESL) method, which combines the time efficiency of gradient calculation of the ESL method with the higher accuracy of gradient calculation in the fully coupled methods. This modified approach, which takes into account the linear dependencies of inertial loads on acceleration, is tested on the application example of a flexible slider-crank mechanism, and the results are compared with the weakly coupled ESL method and a fully coupled optimization where gradients are calculated using the adjoint variable method. [ABSTRACT FROM AUTHOR]

In order to reduce the cost of oceanographic exploration, a new underwater vehicle is designed to sail the required distance with the lowest energy consumed. Since the new underwater vehicle is a complicated multidisciplinary system, it is firstly decomposed into four smaller disciplines and then a multidisciplinary design optimization (MDO) problem is built based on these disciplines. The Multidisciplinary Feasible (MDF) architecture is adopted as the solution strategy to this optimization problem considering that it is easily implemented and a multidisciplinary feasible solution is always guaranteed throughout the optimization process. To solve this optimization problem efficiently, the coupled adjoint method is firstly introduced to improve the efficiency of gradient calculation and then a discipline-merging method is proposed to further enhance the computational efficiency. After this, the discipline-merging method is verified against the finite difference method in two aspects of solution accuracy and computational costs and the results show it is an effective and high efficient gradient calculation method. Finally, the multidisciplinary design optimization of the new underwater vehicle is performed efficiently under the MDF architecture with the discipline-merging method to calculate gradients. [ABSTRACT FROM AUTHOR]

For the issues of magnetic resonance images segmentation, this paper proposed a method based on histogram gradient calculation. Firstly, it smoothed the histogram of MR image so as to remove the lowest grayscale of three individuals. Then, it calculated the gradient on the preprocessed histogram. Finally, it segmented the image after calculating successfully number of objects and their gradients in which they lay. The proposed method was purely based on histogram processing for gradient calculation, so the computational complexity was reduced greatly. Experimental results on T2 weighted MR brain images show that the primary brain areas are extracted out efficiently from 2D and 3D images by the proposed method which has been successfully implemented on human brain MR images obtained in clinical environment. It has better segmentation efficiency than the several existing segmentation algorithms. [ABSTRACT FROM AUTHOR]

Aiming at the defect that the traditional Canny operator edge detection easily lost the edge de-tails,an improved technology was developed to extract the edges of the leaf disease region.More complete gradient information was preserved by using Otsu to binary plant disease images and adding on two oblique directions gradient information extraction,so as to get a more complete edge image.The experimental re-sults showed that this method could effectively filter out noise in the image and detect the details of the gra-dient,remove the false edges and noise edges and get the lesion edges more accurately. [ABSTRACT FROM AUTHOR]

In a previous paper, we described the adaptation of the generalized hybrid orbital (GHO) method for the use in fragment based quantum chemical approaches for large molecules like the field-adapted adjustable density matrix assembler (FA-ADMA). It soon became evident that the limitations of the available implementation strongly restrict its use. Therefore, in the article presented here, we describe new developments to circumvent these limitations. These include new construction algorithms of the GHO hybrid orbitals for sp3- as well as sp2-hybridized carbon and nitrogen, different distributions of electrons in the auxiliary orbitals, and the separation from the CHARMM program, which was needed to provide partial charges and bonding information. The results demonstrate an excellent improvement in accuracy of the obtained electron densities compared to capping hydrogen atoms. Additionally, the calculation of gradients opens the possibility to use FA-ADMA for geometry optimization of large molecules. © 2009 Wiley Periodicals, Inc. Int J Quantum Chem, 2009 [ABSTRACT FROM AUTHOR]

Applying the classical chi-square fitting procedure for multiparameter systems is in somecases extremely difficult due to the lack of an analytical expression for the theoretical functionsdescribing the system. This paper presents an analysis procedure for experimental datausing theoretical functions generated by Monte Carlo method, each corresponding to definitevalues of the minimization parameters. It was applied for the E742 experiment (TRIUMF,Vancouver, Canada) data analysis with the aim to analyze data from Muon Catalyzed Fusionexperiments (extraction muonic atom scattering parameters and parameters of pd fusion inpdμ molecule).