检索结果-内蒙古大学图书馆

On the lactone content distribution estimation in Andrographis paniculata (burm.f.) *** nees using Hyperspectral Images and U-Net Network 20

On the lactone content distribution estimation in Andrograph...

引用

20th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology, ECTI-CON 2023

作者： Rojrattanatrai, N. Kasetkasem, T. Phatrapornant, T. Theerawitaya, C. Chungloo, D. Cha-Um, S. Yamaguchi, M. Kasetsart University Faculty of Engineering Department of Electrical Engineering Bangkok Thailand National Science and Technology Development Agency Pathum Thani Thailand Tokyo Institute of Technology Tokyo Japan

ISBN: (纸本)9798350310467

Hyperspectral reflectance data in the VNIR-SWIR range (400-2500nm) are commonly used to non-destructively and contactless measure the chemical composition of the plants. Most traditional methods are based on non-spatial analysis, that method required the average spectral data to represent the entire image, resulting in the loss of spatial information. To address this issue, we utilize a U-Net network to preserve spatial information while also allowing for the identification and quantification of lactone content in the image. The resulting distribution map provides a clear visualization of lactone content throughout the field or crop, making it easy to identify areas with high or low lactone levels. The pre-processing method includes image registration, outlier removal, spectral smoothing, and normalization. These steps are designed to correct errors and improve the quality of the image and masking. According to the experimental results, the U-Net model achieved R2, RMSE, and DICE of 0.718, 11.66, and 89.92%, respectively. The results show that using hyperspectral images combined with the U-Net network can perform a reliable and accurate prediction model for determining lactone content in A. paniculata. © 2023 ieee.

关键词： Esters

来源：评论

学校读者我要写书评

暂无评论

How Does Visualisation Help App Practitioners Analyse Android Apps?

引用

ieee TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING 2023年第3期20卷 2238-2255页

作者： Tang, Lihong Wu, Tingmin Chen, Xiao Wen, Sheng Li, Li Xia, Xin Grobler, Marthie Xiang, Yang Swinburne Univ Technol Dept Comp Technol Melbourne Vic 3122 Australia CSIROs Data61 Docklands Vic 3008 Australia Monash Univ Fac Informat Technol Clayton Vic 3800 Australia Swinburne Univ Technol Dept Comp Technol Melbourne Vic 3122 Australia Software Engn Applicat Technol Lab Huawei 310027 Peoples R China

Behaviour analysis is essential for the security verification of suspicious Android applications, but analysts are usually faced with a huge obstacle when conducting the app behaviour analysis. They are expected to have comprehensive knowledge of different IT fields and a strong awareness of cyber threats. However, training a new security analyst typically requires a significant amount of time and can be extremely costly. Although there are tools available to assist analysts in studying Android behaviour and security, the completion of this task still heavily relies on the experience of the analysts. To address this problem, we recognise visualisation as a promising method and conduct a series of controlled experiments to demonstrate its effectiveness in the context of Android app behaviour and security analysis. We accordingly develop a visualisation tool based on apps' call graphs (CG) (named VisualDroid) and conduct an experiment and a follow-up interview. Compared to existing solutions, the results suggest that the CG-based visualisation solution (VisualDroid) can lower the barriers to Android behaviour and security analysis. The user study reveals that the platform includes CG-based visualisation components leads to a statistically significant improvement in Android behaviour analysis and security awareness. More specifically, it improves APK Analyzer, JD-GUI, JD-GUI+FlowDroid by 71.4%, 35.7%, and 39.2% in terms of the effectiveness of behaviour analysis. Participants who use VisualDroid also show improvements in the aspect of security awareness with an increase of 155% against APK Analyzer, 96% against JD-GUI, and 59.3% JD-GUI+FlowDroid.

关键词： Codes visualization Security Manuals Inspection Companies Training Android visualisation application comprehension human-computer interaction

来源：评论

学校读者我要写书评

暂无评论

Dual-Affinity Style Embedding Network for Semantic-Aligned Image Style Transfer

引用

ieee TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2023年第10期34卷 7404-7417页

作者： Ma, Zhuoqi Lin, Tianwei Li, Xin Li, Fu He, Dongliang Ding, Errui Wang, Nannan Gao, Xinbo Xidian Univ Sch Comp Sci & Technol Xian Key Lab Big Data & Intelligent Vis Xian 710071 Peoples R China Baidu Inc Dept Comp Vis Vis Technol Beijing 100080 Peoples R China Xidian Univ Sch Telecommun Engn State Key Lab Integrated Serv Networks Xian 710071 Peoples R China Xidian Univ Sch Elect Engn Xian 710071 Peoples R China Chongqing Univ Posts & Telecommun Chongqing Key Lab Image Cognit Chongqing 400065 Peoples R China

Image style transfer aims at synthesizing an image with the content from one image and the style from another. User studies have revealed that the semantic correspondence between style and content greatly affects subjective perception of style transfer results. While current studies have made great progress in improving the visual quality of stylized images, most methods directly transfer global style statistics without considering semantic alignment. Current semantic style transfer approaches still work in an iterative optimization fashion, which is impractically computationally expensive. Addressing these issues, we introduce a novel dual-affinity style embedding network (DaseNet) to synthesize images with style aligned at semantic region granularity. In the dual-affinity module, feature correlation and semantic correspondence between content and style images are modeled jointly for embedding local style patterns according to semantic distribution. Furthermore, the semantic-weighted style loss and the region-consistency loss are introduced to ensure semantic alignment and content preservation. With the end-to-end network architecture, DaseNet can well balance visual quality and inference efficiency for semantic style transfer. Experimental results on different scene categories have demonstrated the effectiveness of the proposed method.

关键词： Semantics Optimization Feature extraction visualization Correlation Training Real-time systems Dual-affinity semantic style transfer style embedding

来源：评论

学校读者我要写书评

暂无评论

Show, Tell and Rephrase: Diverse Video Captioning via Two-Stage Progressive Training

引用

ieee TRANSACTIONS ON MULTIMEDIA 2023年 25卷 7894-7905页

作者： Liu, Zhu Wang, Teng Zhang, Jinrui Zheng, Feng Jiang, Wenhao Lu, Ke Southern Univ Sci & Technol Dept Comp Sci & Engn Shenzhen 518055 Peoples R China Tsinghua Univ Dept Humanities Beijing 100190 Peoples R China Univ Hong Kong Dept Comp Sci Hong Kong Peoples R China Southern Univ Sci & Technol Res Inst Trustworthy Autonomous Syst Shenzhen 518055 Peoples R China Tencent Data Platform Shenzhen 518001 Peoples R China Univ Chinese Acad Sci Sch Engn Sci Beijing 100049 Peoples R China Peng Cheng Lab Shenzhen 518055 Peoples R China

Describing a video using natural language is an inherently one-to-many translation task. To generate diverse captions, existing VAE-based generative models typically learn factorized latent codes via one-stage training merely from stand-alone video-caption pairs. However, such a paradigm neglects set-level relationships among captions from the same video, not fully capturing the underlying multimodality of the generative process. To overcome this shortcoming, we leverage neighbouring descriptions for the same video that are articulated with noticeable topics and language variations (i.e., paraphrases). To this end, we propose a novel progressive training method by decomposing the learning of latent variables into two stages that are topic-oriented and paraphrase-oriented, respectively. Specifically, the model learns from divergent topic sentences obtained by semantic-based clustering in the first stage. It is then trained again through paraphrases with a cluster-aware adaptive regularization, allowing more intra-cluster variations. Furthermore, we introduce an overall metric DAUM, a Diversity-Accuracy Unified Metric to consider both the precision of the generated caption set and its coverage on the reference set, which has proved to have a higher correlation with human judgment than previous precision-only metrics. Extensive experiments on three large-scale video datasets show that the proposed training strategy can achieve superior performance in terms of accuracy, diversity, and DAUM over several baselines.

关键词： Measurement Training visualization Task analysis Computer science Decoding Electronic mail Diverse video captioning conditional variational autoencoders progressive training

来源：评论

学校读者我要写书评

暂无评论

First- And Third-Person Video Co-Analysis By Learning Spatial-Temporal Joint Attention

引用

ieee TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2023年第6期45卷 6631-6646页

作者： Yu, Huangyue Cai, Minjie Liu, Yunfei Lu, Feng Beihang Univ Sch Comp Sci & Engn State Key Lab Virtual Real Technol & Syst Beijing 100191 Peoples R China Hunan Univ Coll Comp Sci & Elect Engn Changsha 410082 Peoples R China Beihang Univ Beijing Adv Innovat Cter Big Data Based Precis Med Beijing 100191 Peoples R China Peng Cheng Lab Shenzhen 518055 Peoples R China

Recent years have witnessed a tremendous increase of first-person videos captured by wearable devices. Such videos record information from different perspectives than the traditional third-person view, and thus show a wide range of potential usages. However, techniques for analyzing videos from different views can be fundamentally different, not to mention co-analyzing on both views to explore the shared information. In this paper, we take the challenge of cross-view video co-analysis and deliver a novel learning-based method. At the core of our method is the notion of "joint attention", indicating the shared attention regions that link the corresponding views, and eventually guide the shared representation learning across views. To this end, we propose a multi-branch deep network, which extracts cross-view joint attention and shared representation from static frames with spatial constraints, in a self-supervised and simultaneous manner. In addition, by incorporating the temporal transition model of the joint attention, we obtain spatial-temporal joint attention that can robustly capture the essential information extending through time. Our method outperforms the state-of-the-art on the standard cross-view video matching tasks on public datasets. Furthermore, we demonstrate how the learnt joint information can benefit various applications through a set of qualitative and quantitative experiments.

关键词： Cameras Task analysis Feature extraction visualization Robot vision systems Standards Egocentric perception joint attention shared representation first-person video third-person video co-analysis deep learning

来源：评论

学校读者我要写书评

暂无评论

MGML: Multigranularity Multilevel Feature Ensemble Network for Remote Sensing Scene Classification

引用

ieee TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2023年第5期34卷 2308-2322页

作者： Zhao, Qi Lyu, Shuchang Li, Yuewen Ma, Yujing Chen, Lijiang Beihang Univ Dept Elect & Informat Engn Beijing 100191 Peoples R China

Remote sensing (RS) scene classification is a challenging task to predict scene categories of RS images. RS images have two main issues: large intraclass variance caused by large resolution variance and confusing information from large geographic covering area. To ease the negative influence from the above two issues. We propose a multigranularity multilevel feature ensemble network (MGML-FENet) to efficiently tackle the RS scene classification task in this article. Specifically, we propose multigranularity multilevel feature fusion branch (MGML-FFB) to extract multigranularity features in different levels of network by channel-separate feature generator (CS-FG). To avoid the interference from confusing information, we propose a multigranularity multilevel feature ensemble module (MGML-FEM), which can provide diverse predictions by full-channel feature generator (FC-FG). Compared to previous methods, our proposed networks have the ability to use structure information and abundant fine-grained features. Furthermore, through the ensemble learning method, our proposed MGML-FENets can obtain more convincing final predictions. Extensive classification experiments on multiple RS datasets (AID, NWPU-RESISC45, UC-Merced, and VGoogle) demonstrate that our proposed networks achieve better performance than previous state-of-the-art (SOTA) networks. The visualization analysis also shows the good interpretability of MGML-FENet.

关键词： Feature extraction Task analysis Learning systems Fuses Generators data mining Remote sensing Channel-separate feature generation (CS-FG) feature ensemble network full-channel feature generation (FC-FG) multigranularity multilevel feature representation remote sensing (RS) scene classification

来源：评论

学校读者我要写书评

暂无评论

Self-Supervised Temporal Graph Learning based on Multi-Head Self-Attention Weighted Neighborhood Sequences 6

Self-Supervised Temporal Graph Learning based on Multi-Head ...

引用

6th International Conference on Big data and Artificial Intelligence, BDAI 2023

作者： Cao, Yulong University of Science and Technology of China School of Data Science Hefei China

ISBN: (纸本)9798350303667

Most current graph representation learning models require labeled samples for training, which may involve high costs in practice. Therefore, self-supervised learning has first been widely applied in static graph representation learning. By adding or removing operations on the adjacency matrix of static graphs to generate supervisory signals, it enables label-free learning of graph nodes. However, the majority of graphs in the real world are temporal graphs. Unlike static graphs, temporal graphs are usually organized through continuous time node interaction sequences, rather than adjacency matrices, making it difficult for self-supervised learning methods on static graphs to be directly migrated to temporal graphs. To address this challenge, we propose a novel self-supervised temporal graph learning approach, named Self-Supervised Temporal Graph Learning based on Neighborhood Sequences (SSHPD). Initially, we perform two types of masking operations on the neighborhood sequences of nodes, then model the neighborhood sequences of nodes using a Hawkes process. Specifically, in order to preserve the maximum neighborhood information of nodes during masking, we introduce a multi-head self-attention mechanism to calculate the attention weights between nodes, selecting those neighbor nodes with relatively small weights for masking. Additionally, we also utilize the multi-head self-attention mechanism to perform a weighted sum on the masked neighborhood sequences, thereby generating virtual nodes for each node and applying it to the Hawkes process, to maximize the extraction of node neighborhood information. We conducted experimental evaluations on three real-world datasets, the results of which demonstrate that, compared to the baseline methods, the embeddings learned by the SSHPD model exhibit superior performance in tasks such as node classification and network visualization. © 2023 ieee.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

PACE: data-Driven Virtual Agent Interaction in Dense and Cluttered Environments

引用

ieee TRANSACTIONS ON visualization AND COMPUTER GRAPHICS 2023年第5期29卷 2536-2546页

作者： Mullen Jr, James F. F. Manocha, Dinesh Univ Maryland College Pk MD 20742 USA

We present PACE, a novel method for modifying motion-captured virtual agents to interact with and move throughout dense, cluttered 3D scenes. Our approach changes a given motion sequence of a virtual agent as needed to adjust to the obstacles and objects in the environment. We first take the individual frames of the motion sequence most important for modeling interactions with the scene and pair them with the relevant scene geometry, obstacles, and semantics such that interactions in the agents motion match the affordances of the scene (e.g., standing on a floor or sitting in a chair). We then optimize the motion of the human by directly altering the high-DOF pose at each frame in the motion to better account for the unique geometric constraints of the scene. Our formulation uses novel loss functions that maintain a realistic flow and natural-looking motion. We compare our method with prior motion generating techniques and highlight the benefits of our method with a perceptual study and physical plausibility metrics. Human raters preferred our method over the prior approaches. Specifically, they preferred our method 57.1% of the time versus the state-of-the-art method using existing motions, and 81.0% of the time versus a state-of-the-art motion synthesis method. Additionally, our method performs significantly higher on established physical plausibility and interaction metrics. Specifically, we outperform competing methods by over 1.2% in terms of the non-collision metric and by over 18% in terms of the contact metric. We have integrated our interactive system with Microsoft HoloLens and demonstrate its benefits in real-world indoor scenes. Our project website is available at https://***/pace/

关键词： Three-dimensional displays Measurement Solid modeling Animation Videos Geometry Affordances Embodied agents Virtual humans Human Factors

来源：评论

学校读者我要写书评

暂无评论

A Reinforcement Learning Paradigm of Configuring Visual Enhancement for Object Detection in Underwater Scenes

引用

ieee JOURNAL OF OCEANIC ENGINEERING 2023年第2期48卷 443-461页

作者： Wang, Hao Sun, Shixin Bai, Xiao Wang, Jian Ren, Peng China Univ Petr East China Coll Control Sci & Engn Qingdao 266580 Peoples R China China Univ Petr East China Coll Oceanog & Space Informat Qingdao 266580 Peoples R China Minist Nat Resources Technol Innovat Ctr Maritime Silk Rd Marine Resour Qingdao 266580 Peoples R China Beihang Univ Coll Comp Sci & Engn Beijing Adv Innovat Ctr Big Data & Brain Comp State Key Lab Software Dev Environm Beijing Peoples R China China Univ Petr East China Coll Sci Qingdao 266580 Peoples R China

This article investigates the problem of enhancing underwater visual observations for the purpose of accurate underwater object detection. Most existing underwater visual enhancement algorithms tend to follow human vision preference but do not necessarily favor the effectiveness of an object detection algorithm. We observe that it should not be the human vision preference but the object detection algorithm that knows what underwater visual enhancement configuration is most beneficial to the detection tasks. In light of this observation, we propose a reinforcement learning paradigm of configuring visual enhancement for object detection in underwater scenes. Specifically, we use underwater image features as states and object detection score increments as rewards. We set up a collection of extensible actions that consist of multiple visual enhancement algorithms. The optimal policy is learned in the form of an action sequence, which characterizes a stepwise process of visual enhancement. Experimental results validate that the sequence of visual enhancement algorithms configured with respect to the object detection algorithm is in favor of improving the detection results.

关键词： visualization Object detection Image color analysis Detectors Feature extraction Reinforcement learning Image enhancement reinforcement learning underwater scenes visual enhancement

来源：评论

学校读者我要写书评

暂无评论

MorphText: Deep Morphology Regularized Accurate Arbitrary-Shape Scene Text Detection

引用

ieee TRANSACTIONS ON MULTIMEDIA 2023年 25卷 4199-4212页

作者： Xu, Chengpei Jia, Wenjing Wang, Ruomei Luo, Xiaonan He, Xiangjian Univ Technol Sydney Fac Engn & IT Ultimo NSW 2007 Australia Univ Nottingham Ningbo China Sch Comp Sci Ningbo 315104 Zhejiang Peoples R China Sun Yat sen Univ Natl Engn Res Ctr Digital Life Sch Data & Comp Sci Guangzhou 510275 Guangdong Peoples R China Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China

Bottom-up text detection methods play an important role in arbitrary-shape scene text detection but there are two restrictions preventing them from achieving their great potential, i.e., 1) the accumulation of false text segment detections, which affects subsequent processing, and 2) the difficulty of building reliable connections between text segments. Targeting these two problems, we propose a novel approach, named "MorphText," to capture the regularity of texts by embedding deep morphology for arbitrary-shape text detection. Towards this end, two deep morphological modules are designed to regularize text segments and determine the linkage between them. First, a Deep Morphological Opening (DMOP) module is constructed to remove false text segment detections generated in the feature extraction process. Then, a Deep Morphological Closing (DMCL) module is proposed to allow text instances of various shapes to stretch their morphology along their most significant orientation while deriving their connections. Extensive experiments conducted on four challenging benchmark datasets (CTW1500, Total-Text, MSRA-TD500 and ICDAR2017) demonstrate that our proposed MorphText outperforms both top-down and bottom-up state-of-the-art arbitrary-shape scene text detection approaches.

关键词： Morphology Couplings Shape Morphological operations Image segmentation visualization Feature extraction Arbitrary-shape scene text detection bottom-up methods deep morphology regularized text segments

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：