Geometric Network Research Articles

Real-time Facial Expression Recognition using 3D Appearance and Geometric Network for Public Security

Real-time Facial Expression Recognition using 3D Appearance and Geometric Network for Public Security

Análise morfométrica e caracterização geomorfológica da bacia hidrográfica do rio Jaboatão (BHRJ) – Pernambuco

O delineamento e caracterização morfométrica de uma bacia hidrográfica são elementos fundamentais da dinâmica deste ambiente, possibilitando contribuições no planejamento e gerenciamento dos recursos hídricos. A bacia hidrográfica do rio Jaboatão localiza-se na Mata Sul do estado de Pernambuco, incluída no grupo de pequenas bacias litorâneas. Para sua caracterização morfométrica utilizou-se os preceitos teóricos de Tonello (2005) e os procedimentos metodológicos de Christofoletti, (1980), enquanto para o mapeamento geomorfológico aplicou-se a proposta de Demek (1972). Os produtos alcançados na análise morfométrica foram segmentados em três critérios: geométrico, relevo e rede de drenagem, alguns resultados principais foram: área, perímetro, coeficiente de compacidade, declividade, hierarquia fluvial, densidade de drenagem e comprimento dos canais. Na caracterização geomorfológica foi mapeado três unidades morfoestruturais e seis unidades geomorfológicas para a área de estudo. Os resultados obtidos a partir dos índices morfométricos e do mapeamento geomorfológico, possibilitaram uma análise integrada da paisagem geomorfológica da bacia do rio Jaboatão.

Gum-Net: Unsupervised Geometric Matching for Fast and Accurate 3D Subtomogram Image Alignment and Averaging.

We propose a Geometric unsupervised matching Network (Gum-Net) for finding the geometric correspondence between two images with application to 3D subtomogram alignment and averaging. Subtomogram alignment is the most important task in cryo-electron tomography (cryo-ET), a revolutionary 3D imaging technique for visualizing the molecular organization of unperturbed cellular landscapes in single cells. However, subtomogram alignment and averaging are very challenging due to severe imaging limits such as noise and missing wedge effects. We introduce an end-to-end trainable architecture with three novel modules specifically designed for preserving feature spatial information and propagating feature matching information. The training is performed in a fully unsupervised fashion to optimize a matching metric. No ground truth transformation information nor category-level or instance-level matching supervision information is needed. After systematic assessments on six real and nine simulated datasets, we demonstrate that Gum-Net reduced the alignment error by 40 to 50% and improved the averaging resolution by 10%. Gum-Net also achieved 70 to 110 times speedup in practice with GPU acceleration compared to state-of-the-art subtomogram alignment methods. Our work is the first 3D unsupervised geometric matching method for images of strong transformation variation and high noise level. The training code, trained model, and datasets are available in our open-source software AITom.

GRNet: Geometric relation network for 3D object detection from point clouds

Rapid detection of 3D objects in indoor environments is essential for indoor mapping and modeling, robotic perception and localization, and building reconstruction. 3D point clouds acquired by a low-cost RGB-D camera have become one of the most commonly used data sources for 3D indoor mapping. However, due to the sparse surface, empty object center, and various scales of point cloud objects, 3D bounding boxes are challenging to be estimated and located accurately. To address this, geometric shape, topological structure, and object relation are commonly employed to extract box reasoning information. In this paper, we describe the geometric feature among object points as an intra-object feature and the relation feature between different objects as an inter-object feature. Based on these two features, we propose an end-to-end point cloud geometric relation network focusing on 3D object detection, which is termed as geometric relation network (GRNet). GRNet first extracts intra-object and inter-object features for each representative point using our proposed backbone network. Then, a centralization module with a scalable loss function is proposed to centralize each representative object point to its center. Next, proposal points are sampled from these shifted points, following a proposal feature pooling operation. Finally, an object-relation learning module is applied to predict bounding box parameters. Such parameters are the additive sum of prediction results from the relation-based inter-object feature and the aggregated intra-object feature. Our model achieves state-of-the-art 3D detection results with 59.1% mAP@0.25 and 39.1% mAP@0.5 on ScanNetV2 dataset, 58.4% mAP@0.25 and 34.9% mAP@0.5 on SUN RGB-D dataset.

Prediction and diagnosis of vertebral tumors on the Internet of Medical Things Platform using geometric rough propagation neural network

Vertebral tumors have a percentage of back pain that causes other vertebral region-born symptoms. Cancers that affect the vertebral column are visceral organ cancer metastases that are mostly seen in older patients. Vertebral dysfunction and neurological failure vertebral column cancers are the most important occurred cancers for patients. In the past, only few methods have been used to combat main and metastatic vertebral tumors. These methods are accessible for short-term monitoring and possess standardized classification consistency for vertebral diagnosis. In this paper, geometric rough propagation neural network has been used for the identification of genetic factors in the examination of a clinical sample with vertebral columns. The proposed neural network has C-statistics of 79.1%, a parameter pitch of 96.1%, and configuration for measurement in the study range with the Brier’s score of 95.6%. The algorithm shows great net gain on the decision curve study, with promising performance results of 98.5% on internal testing for preoperative non-routine estimation of discharges with 0.5% error rate and 96% accuracy range. Also, these models have been externally validated by the online healthcare careers cloud-based open access web application on Internet of Medical Things Platform with 97.9% specificity ratio.

Deep Joint Spatiotemporal Network (DJSTN) for Efficient Facial Expression Recognition.

Understanding a person’s feelings is a very important process for the affective computing. People express their emotions in various ways. Among them, facial expression is the most effective way to present human emotional status. We propose efficient deep joint spatiotemporal features for facial expression recognition based on the deep appearance and geometric neural networks. We apply three-dimensional (3D) convolution to extract spatial and temporal features at the same time. For the geometric network, 23 dominant facial landmarks are selected to express the movement of facial muscle through the analysis of energy distribution of whole facial landmarks.We combine these features by the designed joint fusion classifier to complement each other. From the experimental results, we verify the recognition accuracy of 99.21%, 87.88%, and 91.83% for CK+, MMI, and FERA datasets, respectively. Through the comparative analysis, we show that the proposed scheme is able to improve the recognition accuracy by 4% at least.

Deep geometric convolutional network for automatic modulation classification

A recent trend of automatic modulation classification is to automatically learn high-level abstraction of signals, instead of manually designing features for further classification. In this paper, we propose a new deep geometric convolutional network (DGCN) to hierarchically extract discriminative features from Wigner–Ville distribution map of signals. A group of geometric filters are constructed from a least square support vector machine, to capture the linear singularity existed in maps. The filters are cascaded to construct a deep network for extracting discriminative features and classifying signals with different modulation types. Some experiments are taken to investigate the performance of DGCN, and the results show that it can achieve high accuracy in classifying 15 types of modulation signals.

DWT and CNN based multi-class motor imagery electroencephalographic signal recognition

Objective. Brain computer interface (BCI) system allows humans to control external devices through motor imagery (MI) signals. However, many existing feature extraction algorithms cannot eliminate the influence of individual differences. This research proposed a new processing algorithm that can reduce the impact of individual differences on classification and improve the universality of the algorithm. Approach. To select the optimal frequency band, the energy in each sub-band was calculated by the discrete wavelet transform. Power spectral density and visual geometric group network based convolutional neural network were used for feature extraction and classification respectively. Main results. The test of the BCI Competition IV dataset IIa proved the superiority of the algorithm. In comparison with some commonly used methods, the proposed algorithm reduced classification calculation time while improving classification accuracy; the average classification accuracy rate reaches 96.21%, which is far exceeding the results obtained by the latest literature. Significance. The good classification performance of this research was rooted in the reduced number of parameters, the reduced consumption of computing resources, and the eliminated influence of individual differences. Therefore, the proposed algorithm can be applied to a real-time multi-class BCI system.

Improving the Accuracy for Offline Arabic Digit Recognition Using Sliding Window Approach

Handwritten Digit recognition is a challenging problem these days due to the widely used Arabic language in the world, especially in the Middle East region. In this paper, sliding windows are used to enhance classification accuracies and implemented using random forests (RF) and support vector machine (SVM) classifiers for recognition of Arabic digit images. In order to study their effectiveness with and without using sliding windows, four different feature extraction techniques have been proposed which includes Mean-based, Gray-Level Co-occurrence Matrix (GLCM), Moment-based, and Edge Direction Histogram (EDH). The obtained accuracies show the significance of using sliding windows for classifying digit. The recognition rates acquired using the modified version of AHDBase dataset are 98% when Mean-based and Moment-based are applied with RF classifier, 98.33% and 99.13% when GLCM and EDH are used with linear-kernel SVM, respectively. Moreover, the performance of this study is compared against recent state-of-the-art approaches, namely Geometric-based, two-dimensional discrete cosine transform, Hierarchical features, Hetero-features, Discrete Fourier Transform and geometrical features, Gabor-based, gradient, structural, and concavity and Local Binary Convolutional Neural Networks.

Robust Self-Supervised Learning of Deterministic Errors in Single-Plane (Monoplanar) and Dual-Plane (Biplanar) X-Ray Fluoroscopy

Fluoroscopic imaging that captures X-ray images at video framerates is advantageous for guiding catheter insertions by vascular surgeons and interventional radiologists. Visualizing the dynamical movements non-invasively allows complex surgical procedures to be performed with less trauma to the patient. To improve surgical precision, endovascular procedures can benefit from more accurate fluoroscopy data via calibration. This paper presents a robust self-calibration algorithm suitable for single-plane and dual-plane fluoroscopy. A three-dimensional (3D) target field was imaged by the fluoroscope in a strong geometric network configuration. The unknown 3D positions of targets and the fluoroscope pose were estimated simultaneously by maximizing the likelihood of the Student-t probability distribution function. A smoothed k-nearest-neighbour (kNN) regression is then used to model the deterministic component of the image reprojection error of the robust bundle adjustment. The Maximum Likelihood Estimation step and the kNN regression step are then repeated iteratively until convergence. Four different error modeling schemes were compared while varying the quantity of training images. It was found that using a smoothed kNN regression can automatically model the systematic errors in fluoroscopy with similar accuracy as a human expert using a small training dataset. When all training images were used, the 3D mapping error was reduced from 0.61-0.83 mm to 0.04 mm post-calibration (94.2-95.7% improvement), and the 2D reprojection error was reduced from 1.17-1.31 to 0.20-0.21 pixels (83.2-83.8% improvement). When using biplanar fluoroscopy, the 3D measurement accuracy of the system improved from 0.60 mm to 0.32 mm (47.2% improvement).

The task 3 of forming a network on the sphere from the circles of the same radius

one of the methods of formation of geometric networks of arches of the same radius using regular spherical polyhedra is Investigated. The conditions of the task of placing the specified network on the sphere are set. The criterion for evaluating the effectiveness of solving the problem is the minimum number of standard sizes of segments of the dome arches, the possibility of using technologies of enlarging assembly. The solution of one variant of the problem of placing a geometric network on a spherical cube and, accordingly, on a sphere is given. Placement on the sphere of arches of one radius, different from placement in the form of meridians, has an effective solution in the form of a network with minimal dimensions of arch segments and nodes of paired arches formed on the basis of circles of identical radii formed on the basis of regular spherical polyhedra. The problem is solved by constructing an independent framework of arches of the same radius on the basis of paired circles of the same radius in the spherical cube system.

Epidemic Spreading in Geometric Network with Mobile Agents

Epidemic Spreading in Geometric Network with Mobile Agents

GGM-Net: Graph Geometric Moments Convolution Neural Network for Point Cloud Shape Classification

Geometric feature acts as an important role in point cloud shape classification tasks. Previous methods have proved that the geometric information of point clouds effectively improves the classification accuracy. Mo-Net firstly introduced geometric moments into point cloud shape classification, which, to fit the form of second order geometric moments, extends the number of input channels from three to nine. Unfortunately, similar to PointNet, Mo-Net cannot capture the local structures. To address this issue, we propose a graph geometric moments convolution neural network (GGM-Net), which learns local geometric features from geometric moments representation of a local point set. The core module of the GGM-Net is to learn features from geometric moments (termed as GGM convolution). Specifically, the GGM convolution learns point features and local features from the first and second order geometric moments of points and its local neighbors, respectively, and then combines these features by using an addition operation. In this way, a geometrical local representation about points is obtained, which leads to much surface geometry awareness and robustness. Equipped with the GGM convolution, GGM-Net, a simple end-to-end architecture, is developed to achieve a competitive accuracy on the benchmark dataset ModelNet40 and perform more efficiently in terms of memory and computational complexity.

Mean-Field Game Theoretic Edge Caching in Ultra-Dense Networks

This paper investigates a cellular edge caching problem under a very large number of small base stations (SBSs) and users. In this ultra-dense edge caching network (UDCN), conventional caching algorithms are inapplicable as their complexity increases with the number of small base stations (SBSs). Furthermore, the performance of UDCN is highly sensitive to the dynamics of user demand and inter-SBS interference. To overcome such difficulties, we propose a distributed caching algorithm under a stochastic geometric network model, as well as a spatio-temporal user demand model that characterizes the content popularity dynamics. By exploiting mean-field game (MFG) theory, the complexity of the proposed UDCN caching algorithm becomes independent of the number of SBSs. Numerical evaluations validate that the proposed caching algorithm reduces not only the long run average cost of the network but also the redundant cached data respectively by 24% and 42%, compared to a baseline caching algorithm. The simulation results also show that the proposed caching algorithm is robust to imperfect popularity information, while ensuring a low computational complexity.

Edge-based formulation of elastic network models

We present an edge-based framework for the study of geometric elastic network models to model mechanical interactions in physical systems. We use a formulation in the edge space, instead of the usual node-centric approach, to characterise edge fluctuations of geometric networks defined in d- dimensional space and define the edge mechanical embeddedness, an edge mechanical susceptibility measuring the force felt on each edge given a force applied on the whole system. We further show that this formulation can be directly related to the infinitesimal rigidity of the network, which additionally permits three- and four-centre forces to be included in the network description. We exemplify the approach in protein systems, at both the residue and atomistic levels of description.

Mercator: uncovering faithful hyperbolic embeddings of complex networks

We introduce Mercator, a reliable embedding method to map real complex networks into their hyperbolic latent geometry. The method assumes that the structure of networks is well described by the popularity × similarity static geometric network model, which can accommodate arbitrary degree distributions and reproduces many pivotal properties of real networks, including self-similarity patterns. The algorithm mixes machine learning and maximum likelihood (ML) approaches to infer the coordinates of the nodes in the underlying hyperbolic disk with the best matching between the observed network topology and the geometric model. In its fast mode, Mercator uses a model-adjusted machine learning technique performing dimensional reduction to produce a fast and accurate map, whose quality already outperforms other embedding algorithms in the literature. In the refined Mercator mode, the fast mode embedding result is taken as an initial condition in a ML estimation, which significantly improves the quality of the final embedding. Apart from its accuracy as an embedding tool, Mercator has the clear advantage of systematically inferring not only node orderings, or angular positions, but also the hidden degrees and global model parameters, and has the ability to embed networks with arbitrary degree distributions. Overall, our results suggest that mixing machine learning and ML techniques in a model-dependent framework can boost the meaningful mapping of complex networks.

The geometric dilation of three points

Given three points in the plane, we construct the plane geometric network of smallest geometric dilation that connects them. The geometric dilation of a plane network is defined as the maximum dilation (distance along the network divided by Euclidean distance) between any two points on its edges. We show that the optimum network is either a line segment, a Steiner tree, or a curve consisting of two straight edges and a segment of a logarithmic spiral.

Pointwise geometric and semantic learning network on 3D point clouds

Pointwise geometric and semantic learning network on 3D point clouds

High-order couplings in geometric complex networks of neurons.

We explore the consequences of introducing higher-order interactions in a geometric complex network of Morris-Lecar neurons. We focus on the regime where traveling synchronization waves are observed from a first-neighbors-based coupling to evaluate the changes induced when higher-order dynamical interactions are included. We observe that the traveling-wave phenomenon gets enhanced by these interactions, allowing the activity to travel further in the system without generating pathological full synchronization states. This scheme could be a step toward a simple phenomenological modelization of neuroglial networks.

OPTIMIZATION OF EMERGENCY EVACUATION IN FIRE BUILDING BY INTEGRATED BIM AND GIS

Abstract. Building Information Model (BIM) is a database, which makes a detailed 3D geometrical model with rich semantic information of building, go beyond the standard Computer-Aided Design approach. BIM could be as an ideal source to store, formation model and analyses spatial information of internal environments. GIS is an effective tool for analysing the constant and dynamic variables in small areas such as indoor environments. Therefore, integration of the BIM and GIS is useful for integrating the internal environment and networks for designing optimal routes in emergency evacuation problems. Increasing complexity of high-rise buildings and underground structures lead to much more complication in urban disaster management. One of the main challenges in disaster management and emergency evacuation is to guide people in optimal routes to reach safe areas. In this paper, in order to create optimal routes in internal environments and create an internal geometric network model based on the internal nodes and edges, BIM of a 3D commercial building located in the 8th district of Tehran, was made and transferred to the ARCGIS that enables us to use routing algorithms to find the optimize routes between both internal positions. Flammability, occupancy load, width of exit access and distance of fire point criteria were considered to find the safest routes in emergency evacuation times. Network Analysis and Ant Colony's algorithm were applied for finding the safest routes. The results indicated that the Network Analysis was better in terms of processing time and finding short and safe paths related to the Ant Colony's algorithm.