Search

Your search keyword '"Wang Hongcheng"' showing total 996 results
Start Over Author "Wang Hongcheng"
996 results on '"Wang Hongcheng"'

51. Multilayer Dense Connections for Hierarchical Concept Classification

Author

Parag, Toufiq and Wang, Hongcheng

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Classification is a pivotal function for many computer vision tasks such as object classification, detection, scene segmentation. Multinomial logistic regression with a single final layer of dense connections has become the ubiquitous technique for CNN-based classification. While these classifiers project a mapping between the input and a set of output category classes, they do not typically yield a comprehensive description of the category. In particular, when a CNN based image classifier correctly identifies the image of a Chimpanzee, its output does not clarify that Chimpanzee is a member of Primate, Mammal, Chordate families and a living thing. We propose a multilayer dense connectivity for concurrent prediction of category and its conceptual superclasses in hierarchical order by the same CNN. We experimentally demonstrate that our proposed network can simultaneously predict both the coarse superclasses and finer categories better than several existing algorithms in multiple datasets.

Published
2020

52. VideoSSL: Semi-Supervised Learning for Video Classification

Author

Jing, Longlong, Parag, Toufiq, Wu, Zhe, Tian, Yingli, and Wang, Hongcheng

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

We propose a semi-supervised learning approach for video classification, VideoSSL, using convolutional neural networks (CNN). Like other computer vision tasks, existing supervised video classification methods demand a large amount of labeled data to attain good performance. However, annotation of a large dataset is expensive and time consuming. To minimize the dependence on a large annotated dataset, our proposed semi-supervised method trains from a small number of labeled examples and exploits two regulatory signals from unlabeled data. The first signal is the pseudo-labels of unlabeled examples computed from the confidences of the CNN being trained. The other is the normalized probabilities, as predicted by an image classifier CNN, that captures the information about appearances of the interesting objects in the video. We show that, under the supervision of these guiding signals from unlabeled examples, a video classification CNN can achieve impressive performances utilizing a small fraction of annotated examples on three publicly available datasets: UCF101, HMDB51 and Kinetics.

Published
2020

53. ZoomCount: A Zooming Mechanism for Crowd Counting in Static Images

Author

Sajid, Usman, Sajid, Hasan, Wang, Hongcheng, and Wang, Guanghui

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

This paper proposes a novel approach for crowd counting in low to high density scenarios in static images. Current approaches cannot handle huge crowd diversity well and thus perform poorly in extreme cases, where the crowd density in different regions of an image is either too low or too high, leading to crowd underestimation or overestimation. The proposed solution is based on the observation that detecting and handling such extreme cases in a specialized way leads to better crowd estimation. Additionally, existing methods find it hard to differentiate between the actual crowd and the cluttered background regions, resulting in further count overestimation. To address these issues, we propose a simple yet effective modular approach, where an input image is first subdivided into fixed-size patches and then fed to a four-way classification module labeling each image patch as low, medium, high-dense or no-crowd. This module also provides a count for each label, which is then analyzed via a specifically devised novel decision module to decide whether the image belongs to any of the two extreme cases (very low or very high density) or a normal case. Images, specified as high- or low-density extreme or a normal case, pass through dedicated zooming or normal patch-making blocks respectively before routing to the regressor in the form of fixed-size patches for crowd estimate. Extensive experimental evaluations demonstrate that the proposed approach outperforms the state-of-the-art methods on four benchmarks under most of the evaluation criteria.

Published
2020

63. AmbuBox: A Fast-Deployable Low-Cost Ventilator for COVID-19 Emergent Care

Author

Fang, Zecong, Li, Andrew I, Wang, Hongcheng, Zhang, Ruoyu, Mai, Xiyan, and Pan, Tingrui

Subjects
Manufacturing Engineering, Engineering, Infectious Diseases, Emerging Infectious Diseases, Coronaviruses, Good Health and Well Being, COVID-19, Costs and Cost Analysis, Emergency Medical Services, Equipment Design, Humans, Manufacturing and Industrial Facilities, Pandemics, Respiration, Artificial, SARS-CoV-2, Ventilators, Mechanical, ventilator, AmbuBag, bag valve mask, pneumatic, Medical biotechnology, Biomedical engineering
Abstract

We present a low-cost clinically viable ventilator design, AmbuBox, using a controllable pneumatic enclosure and standard manual resuscitators that are readily available (AmbuBag), which can be rapidly deployed during pandemic and mass-casualty events with a minimal set of components to manufacture and assemble. The AmbuBox is designed to address the existing challenges presented in the existing low-cost ventilator designs by offering an easy-to-install and simple-to-operate apparatus while maintaining a long lifespan with high-precision flow control. As an outcome, a mass-producible prototype of the AmbuBox has been devised, characterized, and validated in a bench test setup using a lung simulator. This prototype will be further investigated through clinical testing. Given the potentially urgent need for inexpensive and rapidly deployable ventilators globally, the overall design, operational principle, and device characterization of the AmbuBox system have been described in detail with open access online. Moreover, the fabrication and assembly methods have been incorporated to enable short-term producibility by a generic local manufacturing facility. In addition, a full list of all components used in the AmbuBox has been included to reflect its low-cost nature.

Published
2020

69. Autonomous Navigation System for Indoor Mobile Robots Based on a Multi-sensor Fusion Technology

Author

Wang, Hongcheng, Chen, Niansheng, Yang, Dingyu, Fan, Guangyu, Filipe, Joaquim, Editorial Board Member, Ghosh, Ashish, Editorial Board Member, Prates, Raquel Oliveira, Editorial Board Member, Zhou, Lizhu, Editorial Board Member, Sun, Yuqing, editor, Lu, Tun, editor, Cao, Buqing, editor, Fan, Hongfei, editor, Liu, Dongning, editor, Du, Bowen, editor, and Gao, Liping, editor

Published
2022
Full Text
View/download PDF

71. Layout-induced Video Representation for Recognizing Agent-in-Place Actions

Author

Yu, Ruichi, Wang, Hongcheng, Li, Ang, Zheng, Jingxiao, Morariu, Vlad I., and Davis, Larry S.

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

We address the recognition of agent-in-place actions, which are associated with agents who perform them and places where they occur, in the context of outdoor home surveillance. We introduce a representation of the geometry and topology of scene layouts so that a network can generalize from the layouts observed in the training set to unseen layouts in the test set. This Layout-Induced Video Representation (LIVR) abstracts away low-level appearance variance and encodes geometric and topological relationships of places in a specific scene layout. LIVR partitions the semantic features of a video clip into different places to force the network to learn place-based feature descriptions; to predict the confidence of each action, LIVR aggregates features from the place associated with an action and its adjacent places on the scene layout. We introduce the Agent-in-Place Action dataset to show that our method allows neural network models to generalize significantly better to unseen scenes.

Published
2018

79. ReMotENet: Efficient Relevant Motion Event Detection for Large-scale Home Surveillance Videos

Author

Yu, Ruichi, Wang, Hongcheng, and Davis, Larry S.

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

This paper addresses the problem of detecting relevant motion caused by objects of interest (e.g., person and vehicles) in large scale home surveillance videos. The traditional method usually consists of two separate steps, i.e., detecting moving objects with background subtraction running on the camera, and filtering out nuisance motion events (e.g., trees, cloud, shadow, rain/snow, flag) with deep learning based object detection and tracking running on cloud. The method is extremely slow and therefore not cost effective, and does not fully leverage the spatial-temporal redundancies with a pre-trained off-the-shelf object detector. To dramatically speedup relevant motion event detection and improve its performance, we propose a novel network for relevant motion event detection, ReMotENet, which is a unified, end-to-end data-driven method using spatial-temporal attention-based 3D ConvNets to jointly model the appearance and motion of objects-of-interest in a video. ReMotENet parses an entire video clip in one forward pass of a neural network to achieve significant speedup. Meanwhile, it exploits the properties of home surveillance videos, e.g., relevant motion is sparse both spatially and temporally, and enhances 3D ConvNets with a spatial-temporal attention model and reference-frame subtraction to encourage the network to focus on the relevant moving objects. Experiments demonstrate that our method can achieve comparable or event better performance than the object detection based method but with three to four orders of magnitude speedup (up to 20k times) on GPU devices. Our network is efficient, compact and light-weight. It can detect relevant motion on a 15s surveillance video clip within 4-8 milliseconds on a GPU and a fraction of second (0.17-0.39) on a CPU with a model size of less than 1MB., Comment: WACV18

Published
2018

84. LiDAR-Based Negative Obstacle Detection for Unmanned Ground Vehicles in Orchards.

Author

Xie, Peng, Wang, Hongcheng, Huang, Yexian, Gao, Qiang, Bai, Zihao, Zhang, Linan, and Ye, Yunxiang

Subjects
*TECHNOLOGICAL innovations, *POINT cloud, *AUTONOMOUS vehicles, *ORCHARDS, *LIDAR
Abstract

In orchard environments, negative obstacles such as ditches and potholes pose significant safety risks to robots working within them. This paper proposes a negative obstacle detection method based on LiDAR tilt mounting. With the LiDAR tilted at 40°, the blind spot is reduced from 3 m to 0.21 m, and the ground point cloud density is increased by an order of magnitude. Based on geometric features of laser point clouds (such as rear wall height and density, and spacing jump between points), a method for detecting negative obstacles is presented. This method establishes a mathematical model by analyzing changes in point cloud height, density, and point spacing, integrating features captured from multiple frames to enhance detection accuracy. Experiments demonstrate that this approach effectively detects negative obstacles in orchard environments, achieving a success rate of 92.7% in obstacle detection. The maximum detection distance reaches approximately 8.0 m, significantly mitigating threats posed to robots by negative obstacles in orchards. This research contributes valuable technological advancements for future orchard automation. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

85. Comparative Analysis of Ca 2 + /Cation Antiporter Gene Family in Rosa roxburghii and Enhanced Calcium Stress Tolerance via Heterologous Expression of RrCAX1a in Tobacco.

Author

Zeng, Tuo, Zhu, Liyong, Su, Wenwen, Gu, Lei, Wang, Hongcheng, Du, Xuye, Zhu, Bin, Wang, Caiyun, and Wu, Di

Subjects
PLANT plasma membranes, GENE expression, GENE families, PROMOTERS (Genetics), VITAMIN C
Abstract

Rosa roxburghii, a calciphilic species native to the mountainous regions of Southwest China, is renowned for its high vitamin C and bioactive components, making it valuable for culinary and medicinal uses. This species exhibits remarkable tolerance to the high-calcium conditions typical of karst terrains. However, the underlying mechanisms of this calcium resilience remain unclear. The Ca2+/cation antiporter (CaCA) superfamily plays a vital role in the transport of Ca2+ and other cations and is crucial for plant tolerance to metal stress. However, the roles and evolutionary significance of the CaCA superfamily members in R. roxburghii remain poorly understood. This study identified 22 CaCA superfamily genes in R. roxburghii, categorized into four subfamilies. The gene structures of these RrCaCAs show considerable conservation across related species. Selection pressure analysis revealed that all RrCaCAs are subject to purifying selection. The promoter regions of these genes contain numerous hormone-responsive and stress-related elements. qRT-PCR analyses demonstrated that H+/cation exchanger (CAX) RrCAX1a and RrCAX3a were highly responsive to Ca2+ stress, cation/Ca2+ exchanger (CCX) RrCCX4 to Mg2+ stress, and RrCCX11a to Na+ stress. Subcellular localization indicated that RrCAX1a is localized to the plant cell membrane, and its stable transformation in tobacco confirmed its ability to confer enhanced resistance to heavy Ca2+ stresses, highlighting its crucial role in the high-calcium tolerance mechanisms of R. roxburghii. This research establishes a foundation for further molecular-level functional analyses of the adaptation mechanisms of R. roxburghii to high-calcium environments. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

86. Simultaneously Possessing High Quantum Efficiency and Thermal Robustness in a Near‐Infrared Emitting ZnAlB(1‐x)GaxO4:Cr3+ Phosphor.

Author

Lim, Jin Xiang, Wang, Hongcheng, Kang, Xiaojiao, Lü, Wei, Lim, Chin Seong, Zhu, Zhennan, Pan, Qiwen, Sakundarini, Novita, and Chuah, Yea Dat

Subjects
*QUANTUM efficiency, *NIGHT vision, *THERMAL efficiency, *ENERGY transfer, *IMAGE analysis, *PHOTOTHERMAL effect
Abstract

Near‐infrared (NIR) phosphors are a vital component of phosphor‐converted light‐emitting diodes (pc‐LEDs), with simultaneous high quantum efficiency and thermal stability being essential attributes for optimal performance. In this paper, a novel ZnAlB(1‐x)GaxO4:0.05Cr3+ system is designed by gradually substituting B with Ga. Surprisingly, a significant enhancement in photoluminescence performance is observed, reaching peak efficiency at x = 1. Further development led to a series of ZnAlGaO4:yCr3+ (ZAGO:yCr3+) phosphors, exhibiting variable full width at half maximum (FWHM) ranging from 32 to 163 nm with increasing Cr3+ content, attributed to energy transfer between two Cr3+ emission centers. Among these, ZAGO:0.05Cr3+ emerges as the optimal phosphor, boasting a high internal/external quantum efficiency of 80%/34% with a FWHM of 80 nm. Notably, this phosphor demonstrates anti‐thermal quenching behavior (125.53% @ 423 K), with its exceptional thermal stability elucidated through the defect level model. The fabricated NIR pc‐LEDs featuring ZAGO:Cr3+ phosphors show great prospect in multifunctional applications in plant cultivation, night vision, non‐destructive analysis and veins imaging. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

87. Genome-Wide Identification and Analysis of Phospholipase C Gene Family Reveals Orthologs, Co-Expression Networks, and Expression Profiling Under Abiotic Stress in Sorghum bicolor.

Author

Wang, Hongcheng, Yu, Junxing, Zhang, Xingyu, Zeng, Qian, Zeng, Tuo, Gu, Lei, Zhu, Bin, Yu, Feng, and Du, Xuye

Subjects
GENE expression, PHOSPHOLIPASE C, SORGHUM, ABIOTIC stress, GENE families
Abstract

Phospholipase C (PLC) is an essential enzyme involved in lipid signaling pathways crucial for regulating plant growth and responding to environmental stress. In sorghum, 11 PLC genes have been identified, comprising 6 PI-PLCs and 5 NPCs. Through phylogenetic and interspecies collinearity analyses, structural similarities between SbPLCs and ZmPLCs proteins have been observed, with a particularly strong collinearity between SbPLCs and OsPLCs. Promoter function analysis has shown that SbPLCs are significantly enriched under abiotic stress and hormonal stimuli, like ABA, jasmonic acid, drought, high temperature, and salt. Gene co-expression networks, constructed using a weighted gene co-expression network analysis (WGCNA), highlight distinct expression patterns of SbPLC1, SbPLC3a, and SbPLC4 in response to abiotic stress, providing further insights into the expression patterns and interactions of SbPLCs under various environmental stimuli. qRT-PCR results reveal variations in expression levels among most SbPLCs members under different stress conditions (drought, NaCl, NaHCO3), hormone treatments (ABA), and developmental stages, indicating both specific and overlapping expression patterns. This comprehensive analysis offers valuable insights into the roles of SbPLCs in sorghum, shedding light on their specific expression patterns, regulatory elements, and protein interactions across different environmental stimuli and developmental stages. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results