Showing papers by "National University of Defense Technology published in 2020"

Deep Learning for Generic Object Detection: A Survey

Abstract: Object detection, one of the most fundamental and challenging problems in computer vision, seeks to locate object instances from a large number of predefined categories in natural images. Deep learning techniques have emerged as a powerful strategy for learning feature representations directly from data and have led to remarkable breakthroughs in the field of generic object detection. Given this period of rapid evolution, the goal of this paper is to provide a comprehensive survey of the recent achievements in this field brought about by deep learning techniques. More than 300 research contributions are included in this survey, covering many aspects of generic object detection: detection frameworks, object feature representation, object proposal generation, context modeling, training strategies, and evaluation metrics. We finish the survey by identifying promising directions for future research.

RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds

Abstract: We study the problem of efficient semantic segmentation for large-scale 3D point clouds. By relying on expensive sampling techniques or computationally heavy pre/post-processing steps, most existing approaches are only able to be trained and operate over small-scale point clouds. In this paper, we introduce RandLA-Net, an efficient and lightweight neural architecture to directly infer per-point semantics for large-scale point clouds. The key to our approach is to use random point sampling instead of more complex point selection approaches. Although remarkably computation and memory efficient, random sampling can discard key features by chance. To overcome this, we introduce a novel local feature aggregation module to progressively increase the receptive field for each 3D point, thereby effectively preserving geometric details. Extensive experiments show that our RandLA-Net can process 1 million points in a single pass with up to 200x faster than existing approaches. Moreover, our RandLA-Net clearly surpasses state-of-the-art approaches for semantic segmentation on two large-scale benchmarks Semantic3D and SemanticKITTI.

Population flow drives spatio-temporal distribution of COVID-19 in China.

Abstract: Sudden, large-scale and diffuse human migration can amplify localized outbreaks of disease into widespread epidemics1–4. Rapid and accurate tracking of aggregate population flows may therefore be epidemiologically informative. Here we use 11,478,484 counts of mobile phone data from individuals leaving or transiting through the prefecture of Wuhan between 1 January and 24 January 2020 as they moved to 296 prefectures throughout mainland China. First, we document the efficacy of quarantine in ceasing movement. Second, we show that the distribution of population outflow from Wuhan accurately predicts the relative frequency and geographical distribution of infections with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) until 19 February 2020, across mainland China. Third, we develop a spatio-temporal ‘risk source’ model that leverages population flow data (which operationalize the risk that emanates from epidemic epicentres) not only to forecast the distribution of confirmed cases, but also to identify regions that have a high risk of transmission at an early stage. Fourth, we use this risk source model to statistically derive the geographical spread of COVID-19 and the growth pattern based on the population outflow from Wuhan; the model yields a benchmark trend and an index for assessing the risk of community transmission of COVID-19 over time for different locations. This approach can be used by policy-makers in any nation with available data to make rapid and accurate risk assessments and to plan the allocation of limited resources ahead of ongoing outbreaks. Modelling of population flows in China enables the forecasting of the distribution of confirmed cases of COVID-19 and the identification of areas at high risk of SARS-CoV-2 transmission at an early stage.

Characterizing the Propagation of Situational Information in Social Media During COVID-19 Epidemic: A Case Study on Weibo

Abstract: During the ongoing outbreak of coronavirus disease (COVID-19), people use social media to acquire and exchange various types of information at a historic and unprecedented scale. Only the situational information are valuable for the public and authorities to response to the epidemic. Therefore, it is important to identify such situational information and to understand how it is being propagated on social media, so that appropriate information publishing strategies can be informed for the COVID-19 epidemic. This article sought to fill this gap by harnessing Weibo data and natural language processing techniques to classify the COVID-19-related information into seven types of situational information. We found specific features in predicting the reposted amount of each type of information. The results provide data-driven insights into the information need and public attention.

Multiscale Dynamic Graph Convolutional Network for Hyperspectral Image Classification

Abstract: Convolutional neural network (CNN) has demonstrated impressive ability to represent hyperspectral images and to achieve promising results in hyperspectral image classification. However, traditional CNN models can only operate convolution on regular square image regions with fixed size and weights, and thus, they cannot universally adapt to the distinct local regions with various object distributions and geometric appearances. Therefore, their classification performances are still to be improved, especially in class boundaries. To alleviate this shortcoming, we consider employing the recently proposed graph convolutional network (GCN) for hyperspectral image classification, as it can conduct the convolution on arbitrarily structured non-Euclidean data and is applicable to the irregular image regions represented by graph topological information. Different from the commonly used GCN models that work on a fixed graph, we enable the graph to be dynamically updated along with the graph convolution process so that these two steps can be benefited from each other to gradually produce the discriminative embedded features as well as a refined graph. Moreover, to comprehensively deploy the multiscale information inherited by hyperspectral images, we establish multiple input graphs with different neighborhood scales to extensively exploit the diversified spectral–spatial correlations at multiple scales. Therefore, our method is termed multiscale dynamic GCN (MDGCN). The experimental results on three typical benchmark data sets firmly demonstrate the superiority of the proposed MDGCN to other state-of-the-art methods in both qualitative and quantitative aspects.

Preliminary prediction of the basic reproduction number of the Wuhan novel coronavirus 2019-nCoV.

Abstract: Objectives To estimate the basic reproduction number of the Wuhan novel coronavirus (2019-nCoV). Methods Based on the susceptible-exposed-infected-removed (SEIR) compartment model and the assumption that the infectious cases with symptoms occurred before 26 January, 2020 are resulted from free propagation without intervention, we estimate the basic reproduction number of 2019-nCoV according to the reported confirmed cases and suspected cases, as well as the theoretical estimated number of infected cases by other research teams, together with some epidemiological determinants learned from the severe acute respiratory syndrome (SARS). Results The basic reproduction number fall between 2.8 and 3.3 by using the real-time reports on the number of 2019-nCoV-infected cases from People's Daily in China and fall between 3.2 and 3.9 on the basis of the predicted number of infected cases from international colleagues. Conclusions The early transmission ability of 2019-nCoV is close to or slightly higher than SARS. It is a controllable disease with moderate to high transmissibility. Timely and effective control measures are needed to prevent the further transmissions.

Multiple Kernel $k$ k -Means with Incomplete Kernels

Abstract: Multiple kernel clustering (MKC) algorithms optimally combine a group of pre-specified base kernel matrices to improve clustering performance. However, existing MKC algorithms cannot efficiently address the situation where some rows and columns of base kernel matrices are absent. This paper proposes two simple yet effective algorithms to address this issue. Different from existing approaches where incomplete kernel matrices are first imputed and a standard MKC algorithm is applied to the imputed kernel matrices, our first algorithm integrates imputation and clustering into a unified learning procedure. Specifically, we perform multiple kernel clustering directly with the presence of incomplete kernel matrices, which are treated as auxiliary variables to be jointly optimized. Our algorithm does not require that there be at least one complete base kernel matrix over all the samples. Also, it adaptively imputes incomplete kernel matrices and combines them to best serve clustering. Moreover, we further improve this algorithm by encouraging these incomplete kernel matrices to mutually complete each other. The three-step iterative algorithm is designed to solve the resultant optimization problems. After that, we theoretically study the generalization bound of the proposed algorithms. Extensive experiments are conducted on 13 benchmark data sets to compare the proposed algorithms with existing imputation-based methods. Our algorithms consistently achieve superior performance and the improvement becomes more significant with increasing missing ratio, verifying the effectiveness and advantages of the proposed joint imputation and clustering.

Ultrafast fiber lasers mode-locked by two-dimensional materials: review and prospect

Abstract: The year 2019 marks the 10th anniversary of the first report of ultrafast fiber laser mode-locked by graphene. This result has had an important impact on ultrafast laser optics and continues to offer new horizons. Herein, we mainly review the linear and nonlinear photonic properties of two-dimensional (2D) materials, as well as their nonlinear applications in efficient passive mode-locking devices and ultrafast fiber lasers. Initial works and significant progress in this field, as well as new insights and challenges of 2D materials for ultrafast fiber lasers, are reviewed and analyzed.

Attention Receptive Pyramid Network for Ship Detection in SAR Images

Abstract: With the development of deep learning (DL) and synthetic aperture radar (SAR) imaging techniques, SAR automatic target recognition has come to a breakthrough. Numerous algorithms have been proposed and competitive results have been achieved in detecting different targets. However, due to the influence of various sizes and complex background of ships, detecting multiscale ships in SAR images is still challenging. To solve the problems, a novel network, called attention receptive pyramid network (ARPN), is proposed in this article. ARPN is a two-stage detector and designed to improve the performance of detecting multiscale ships in SAR images by enhancing the relationships among nonlocal features and refining information at different feature maps. Specifically, receptive fields block (RFB) and convolutional block attention module (CBAM) are employed and combined reasonably in attention receptive block to build a top-down fine-grained feature pyramid. RFB, composed of several branches of convolutional layers with specifically asymmetric kernel sizes and various dilation rates, is used for grabbing features of ships with large aspect ratios and enhancing local features with their global dependences. CBAM, which consists of channel and spatial attention mechanisms, is utilized to boost significant information and suppress interference caused by surroundings. To evaluate the effectiveness of ARPN, experiments are conducted on SAR Ship Detection Dataset and two large-scene SAR images. The detection results illustrate that competitive performance has been achieved by our method in comparison with several CNN-based algorithms, e.g., Faster-RCNN, RetinaNet, feature pyramid network, YOLOv3, Dense Attention Pyramid Network, Depth-wise Separable Convolutional Neural Network, High-Resolution Ship Detection Network, and Squeeze and Excitation Rank Faster-RCNN.

Assessing spread risk of Wuhan novel coronavirus within and beyond China, January-April 2020: a travel network-based modelling study

Abstract: Objective: To estimate the potential risk and geographic range of Wuhan novel coronavirus (2019-nCoV) spread within and beyond China from January through to April, 2020. Design: Travel network-based modelling study. Setting and participants: General population travelling from Wuhan and other high-risk cities in China. Main outcome measures: Based on de-identified and aggregated mobile phone data, air passenger itinerary data, and case reports, we defined the relative importation risk and internal and international destinations of 2019-nCoV from Wuhan and other high-risk cities in China. Results: The cordon sanitaire of Wuhan is likely to have occurred during the latter stages of peak population numbers leaving the city before Lunar New Year (LNY), with travellers departing into neighbouring cities and other megacities in China, and a high proportion of cases likely travelled with symptoms at the early stage of the outbreak. Should secondary outbreaks occur in 17 high-risk secondary cities, they could contribute to seeding the virus in other highly connected cities within and beyond China after the LNY holiday. We estimated that 59,912 air passengers, of which 834 (95% UI: 478 - 1349) had 2019-nCoV infection, travelled from Wuhan to 382 cities outside of mainland China during the two weeks prior to Wuhan’s lockdown. The majority of these cities were in Asia, but major hubs in Europe, the US and Australia were also prominent, with strong correlation seen between predicted importation risks and reported cases seen. Because significant spread has already occurred, a large number of airline travellers (3.3 million under the scenario of 75% travel reduction from normal volumes) may be required to be screened at origin high-risk cities in China and destinations across the globe for the following three months of February to April, 2020 to effectively limit spread beyond its current extent. Conclusion: Further spread of 2019-nCoV within China and international exportation is likely to occur. All countries, especially vulnerable regions, should be prepared for efforts to contain the 2019-nCoV infection. Keywords: Coronavirus; Epidemiology; Pandemic; Mobile phone; Air travel

Learning Canonical Shape Space for Category-Level 6D Object Pose and Size Estimation

Abstract: We present a novel approach to category-level 6D object pose and size estimation. To tackle intra-class shape variations, we learn canonical shape space (CASS), a unified representation for a large variety of instances of a certain object category. In particular, CASS is modeled as the latent space of a deep generative model of canonical 3D shapes with normalized pose. We train a variational auto-encoder (VAE) for generating 3D point clouds in the canonical space from an RGBD image. The VAE is trained in a cross-category fashion, exploiting the publicly available large 3D shape repositories. Since the 3D point cloud is generated in normalized pose (with actual size), the encoder of the VAE learns view-factorized RGBD embedding. It maps an RGBD image in arbitrary view into a poseindependent 3D shape representation. Object pose is then estimated via contrasting it with a pose-dependent feature of the input RGBD extracted with a separate deep neural networks. We integrate the learning of CASS and pose and size estimation into an end-to-end trainable network, achieving the state-of-the-art performance.

Towards Real-Time Path Planning through Deep Reinforcement Learning for a UAV in Dynamic Environments

Abstract: Path planning remains a challenge for Unmanned Aerial Vehicles (UAVs) in dynamic environments with potential threats. In this paper, we have proposed a Deep Reinforcement Learning (DRL) approach for UAV path planning based on the global situation information. We have chosen the STAGE Scenario software to provide the simulation environment where a situation assessment model is developed with consideration of the UAV survival probability under enemy radar detection and missile attack. We have employed the dueling double deep Q-networks (D3QN) algorithm that takes a set of situation maps as input to approximate the Q-values corresponding to all candidate actions. In addition, the e-greedy strategy is combined with heuristic search rules to select an action. We have demonstrated the performance of the proposed method under both static and dynamic task settings.

Finding key players in complex networks through deep reinforcement learning.

Abstract: Finding an optimal set of nodes, called key players, whose activation (or removal) would maximally enhance (or degrade) certain network functionality, is a fundamental class of problems in network science1,2. Potential applications include network immunization3, epidemic control4, drug design5, and viral marketing6. Due to their general NP-hard nature, those problems typically cannot be solved by exact algorithms with polynomial time complexity7. Many approximate and heuristic strategies have been proposed to deal with specific application scenarios1,2,8-12. Yet, we still lack a unified framework to efficiently solve this class of problems. Here we introduce a deep reinforcement learning framework FINDER, which can be trained purely on small synthetic networks generated by toy models and then applied to a wide spectrum of influencer finding problems. Extensive experiments under various problem settings demonstrate that FINDER significantly outperforms existing methods in terms of solution quality. Moreover, it is several orders of magnitude faster than existing methods for large networks. The presented framework opens up a new direction of using deep learning techniques to understand the organizing principle of complex networks, which enables us to design more robust networks against both attacks and failures.

Estimating the effective reproduction number of the 2019-nCoV in China

Abstract: We estimate the effective reproduction number for 2019-nCoV based on the daily reported cases from China CDC. The results indicate that 2019-nCoV has a higher effective reproduction number than SARS with a comparable fatality rate. Article Summary Line This modeling study indicates that 2019-nCoV has a higher effective reproduction number than SARS with a comparable fatality rate.

Hierarchical Clustering With Hard-Batch Triplet Loss for Person Re-Identification

Abstract: For clustering-guided fully unsupervised person reidentification (re-ID) methods, the quality of pseudo labels generated by clustering directly decides the model performance. In order to improve the quality of pseudo labels in existing methods, we propose the HCT method which combines hierarchical clustering with hard-batch triplet loss. The key idea of HCT is to make full use of the similarity among samples in the target dataset through hierarchical clustering, reduce the influence of hard examples through hard-batch triplet loss, so as to generate high quality pseudo labels and improve model performance. Specifically, (1) we use hierarchical clustering to generate pseudo labels, (2) we use PK sampling in each iteration to generate a new dataset for training, (3) we conduct training with hard-batch triplet loss and evaluate model performance in each iteration. We evaluate our model on Market-1501 and DukeMTMC-reID. Results show that HCT achieves 56.4% mAP on Market-1501 and 50.7% mAP on DukeMTMC-reID which surpasses state-of-the-arts a lot in fully unsupervised re-ID and even better than most unsupervised domain adaptation (UDA) methods which use the labeled source dataset. Code will be released soon on https://github.com/zengkaiwei/HCT

Pythagorean fuzzy MCDM method based on CoCoSo and CRITIC with score function for 5G industry evaluation

Abstract: The 5G industry is of great concern to countries to formulate a major national strategy for 5G planning, promote industrial upgrading, and accelerate their economic and technological modernization. When considering the 5G industry evaluation, the basic issues involve strong uncertainty. Pythagorean fuzzy sets, depicted by membership degree and non-membership degree, are a more resultful means for capturing uncertainty. In this paper, the comparison issue in Pythagorean fuzzy environment is disposed by proposing novel score function. Next, the $$\ominus $$ and $$\oslash $$ operations are defined and their properties are proved. Later, the objective weight is calculated by Criteria Importance Through Inter-criteria Correlation method. Meanwhile, the combined weight is determined by reflecting both subjective weight and the objective weight. Then, the Pythagorean fuzzy decision making algorithm based Combined Compromise Solution is developed. Lastly, the validity of algorithm is expounded by the 5G evaluation issue, along with their sensitivity analysis. The main advantages of proposed algorithm are: (1) have no counterintuitive phenomena; (2) without division or antilogarithm by zero problem; (3) own stronger ability to distinguish alternatives.

Computation-Efficient Offloading and Trajectory Scheduling for Multi-UAV Assisted Mobile Edge Computing

Abstract: The emergence of mobile edge computing (MEC) and unmanned aerial vehicles (UAVs) is of great significance for the prospective development of Internet of Things (IoT). The additional computation capability and extensive network coverage provide energy-limited smart mobile devices (SMDs) with more opportunities to experience diverse intelligent applications. In this paper, a computation efficiency maximization problem is formulated in a multi-UAV assisted MEC system, where both computation bits and energy consumption are considered. Based on the partial computation offloading mode, user association, allocation of central processing unit (CPU) cycle frequency, power and spectrum resources, as well as trajectory scheduling of UAVs are jointly optimized. Due to the non-convexity of the problem and the coupling among variables, we propose an iterative optimization algorithm with double-loop structure to find the optimal solution. Simulation results demonstrate that the proposed algorithm can obtain higher computation efficiency than baseline schemes while guaranteeing the quality of computation service.

Wood-Inspired Anisotropic Cellulose Nanofibril Composite Sponges for Multifunctional Applications.

Abstract: Nanocellulose-based porous materials have been recently considered as ideal candidates in various applications. However, challenges on performances remain owing to the disorderly structure and the limited transport specificity. Herein, wood-inspired composite sponges consisting of cellulose nanofibrils (CNFs) and high-aspect-ratio silver nanowires (AgNWs) were generated with anisotropic properties by the directional freeze-drying. The obtained composite sponges exhibited attractive features, such as an excellent compressive stress of 24.5 kPa, low percolation threshold of 0.1 vol % AgNWs, and high electrical conductivity of 1.52 S/cm. Furthermore, the self-assembled ordered structure in the longitudinal direction and synergistic effect between CNFs and AgNWs benefited the sponge interesting anisotropic electrical conductivity, thermal diffusivity, ultrafast electrically induced heating (<5 s), sensitive pressure sensing (errors <0.26%), and electromagnetic interference (EMI) shielding for special practical demands. This multifunctional material inspired by natural woods is expected to broaden new applications as electronic devices for an intelligent switch or EMI shielding.

MLCVNet: Multi-Level Context VoteNet for 3D Object Detection

Abstract: In this paper, we address the 3D object detection task by capturing multi-level contextual information with the self-attention mechanism and multi-scale feature fusion. Most existing 3D object detection methods recognize objects individually, without giving any consideration on contextual information between these objects. Comparatively, we propose Multi-Level Context VoteNet (MLCVNet) to recognize 3D objects correlatively, building on the state-of-the-art VoteNet. We introduce three context modules into the voting and classifying stages of VoteNet to encode contextual information at different levels. Specifically, a Patch-to-Patch Context (PPC) module is employed to capture contextual information between the point patches, before voting for their corresponding object centroid points. Subsequently, an Object-to-Object Context (OOC) module is incorporated before the proposal and classification stage, to capture the contextual information between object candidates. Finally, a Global Scene Context (GSC) module is designed to learn the global scene context. We demonstrate these by capturing contextual information at patch, object and scene levels. Our method is an effective way to promote detection accuracy, achieving new state-of-the-art detection performance on challenging 3D object detection datasets, i.e., SUN RGBD and ScanNet. We also release our code at https://github.com/NUAAXQ/MLCVNet.

Deep Video Super-Resolution Using HR Optical Flow Estimation

Abstract: Video super-resolution (SR) aims at generating a sequence of high-resolution (HR) frames with plausible and temporally consistent details from their low-resolution (LR) counterparts. The key challenge for video SR lies in the effective exploitation of temporal dependency between consecutive frames. Existing deep learning based methods commonly estimate optical flows between LR frames to provide temporal dependency. However, the resolution conflict between LR optical flows and HR outputs hinders the recovery of fine details. In this paper, we propose an end-to-end video SR network to super-resolve both optical flows and images. Optical flow SR from LR frames provides accurate temporal dependency and ultimately improves video SR performance. Specifically, we first propose an optical flow reconstruction network (OFRnet) to infer HR optical flows in a coarse-to-fine manner. Then, motion compensation is performed using HR optical flows to encode temporal dependency. Finally, compensated LR inputs are fed to a super-resolution network (SRnet) to generate SR results. Extensive experiments have been conducted to demonstrate the effectiveness of HR optical flows for SR performance improvement. Comparative results on the Vid4 and DAVIS-10 datasets show that our network achieves the state-of-the-art performance.

Survey on Unmanned Aerial Vehicle Networks: A Cyber Physical System Perspective

Abstract: Unmanned aerial vehicle (UAV) networks are playing an important role in various areas due to their agility and versatility, which have attracted significant attentions from both the academia and industry in recent years. As an integration of embedded systems with communication devices, computation capabilities and control modules, the UAV network could build a closed loop from data perceiving, information exchanging, decision making to the final execution, which tightly integrates cyber processes into physical devices. Therefore, the UAV network could be considered as a cyber physical system (CPS). Revealing coupling effects among the three interacted CPS components, i.e., communication, computation and control, is envisioned as the key to properly utilize all available resources and hence improve the performance of the UAV network. In this paper, we present a comprehensive survey on UAV networks from a CPS perspective. Firstly, we review the basics and advances of the three CPS components in UAV networks. Then we look inside to investigate how these components contribute to the system performance by classifying UAV networks into three hierarchies, i.e., cell level, system level, and system of system level. Furthermore, the coupling effects among these CPS components are explicitly illustrated, which could be enlightening to deal with the challenges in each individual aspect. New research directions and open issues are discussed at the end of this survey. With this intensive literature review, we intend to provide a novel insight into the state of the art in UAV networks.

Joint Optimization on Trajectory, Altitude, Velocity, and Link Scheduling for Minimum Mission Time in UAV-Aided Data Collection

Abstract: Due to the flexibility in 3-D space and high probability of line-of-sight (LoS) in air-to-ground communications, unmanned aerial vehicles (UAVs) have been considered as means to support energy-efficient data collection. However, in emergency applications, the mission completion time should be main concerns. In this article, we propose a UAV-aided data collection design to gather data from a number of ground users (GUs). The objective is to optimize the UAV’s trajectory, altitude, velocity, and data links with GUs to minimize the total mission time. However, the difficulty lies in that the formulated time minimization problem has mutual effect with trajectory variables. To tackle this issue, we first transform the original problem equivalently to the trajectory length problem and then decompose the problem into three subproblems: 1) altitude optimization; 2) trajectory optimization; and 3) velocity and link scheduling optimization. In the altitude optimization, the aim is to maximize the transmission region of GUs which can benefit trajectory designing; then, in the trajectory optimization, we propose a segment-based trajectory optimization algorithm (STOA) to avoid repeat travel; besides, we also propose a group-based trajectory optimization algorithm (GTOA) in large-scale high-density GU deployment to relieve massive computation introduced by STOA. Then, the velocity and link scheduling optimization is modeled as a mixed-integer nonlinear programming (MINLP) and block coordinate descent (BCD) is employed to solve it. Simulations show that both STOA and GTOA achieve shorter trajectory compared with the existing algorithm and GTOA has less computational complexity; besides, the proposed time minimization design is valid by comparing to the benchmark scheme.

A Reinforcement Learning Approach to Autonomous Decision Making of Intelligent Vehicles on Highways

Abstract: Autonomous decision making is a critical and difficult task for intelligent vehicles in dynamic transportation environments. In this paper, a reinforcement learning approach with value function approximation and feature learning is proposed for autonomous decision making of intelligent vehicles on highways. In the proposed approach, the sequential decision making problem for lane changing and overtaking is modeled as a Markov decision process with multiple goals, including safety, speediness, smoothness, etc. In order to learn optimized policies for autonomous decision-making, a multiobjective approximate policy iteration (MO-API) algorithm is presented. The features for value function approximation are learned in a data-driven way, where sparse kernel-based features or manifold-based features can be constructed based on data samples. Compared with previous RL algorithms such as multiobjective Q-learning, the MO-API approach uses data-driven feature representation for value and policy approximation so that better learning efficiency can be achieved. A highway simulation environment using a 14 degree-of-freedom vehicle dynamics model was established to generate training data and test the performance of different decision-making methods for intelligent vehicles on highways. The results illustrate the advantages of the proposed MO-API method under different traffic conditions. Furthermore, we also tested the learned decision policy on a real autonomous vehicle to implement overtaking decision and control under normal traffic on highways. The experimental results also demonstrate the effectiveness of the proposed method.

PQ-NET: A Generative Part Seq2Seq Network for 3D Shapes

Abstract: We introduce PQ-NET, a deep neural network which represents and generates 3D shapes via sequential part assembly. The input to our network is a 3D shape segmented into parts, where each part is first encoded into a feature representation using a part autoencoder. The core component of PQ-NET is a sequence-to-sequence or Seq2Seq autoencoder which encodes a sequence of part features into a latent vector of fixed size, and the decoder reconstructs the 3D shape, one part at a time, resulting in a sequential assembly. The latent space formed by the Seq2Seq encoder encodes both part structure and fine part geometry. The decoder can be adapted to perform several generative tasks including shape autoencoding, interpolation, novel shape generation, and single-view 3D reconstruction, where the generated shapes are all composed of meaningful parts.