检索结果-内蒙古大学图书馆

Emotional Semantics-Preserved and Feature-Aligned CycleGAN for Visual Emotion Adaptation

ieee TRANSACTIONS ON CYBERNETICS 2022年第10期52卷 10000-10013页

作者： Zhao, Sicheng Chen, Xuanbai Yue, Xiangyu Lin, Chuang Xu, Pengfei Krishna, Ravi Yang, Jufeng Ding, Guiguang Sangiovanni-Vincentelli, Alberto L. Keutzer, Kurt Tsinghua Univ BNRist Beijing 100084 Peoples R China Nankai Univ Coll Comp Sci Tianjin 300350 Peoples R China Univ Calif Berkeley Dept Elect Engn & Comp Sci Berkeley CA 94709 USA Didi Chuxing Beijing 100193 Peoples R China

Thanks to large-scale labeled training data, deep neural networks (DNNs) have obtained remarkable success in many vision and multimedia tasks. However, because of the presence of domain shift, the learned knowledge of the well-trained DNNs cannot be well generalized to new domains or datasets that have few labels. Unsupervised domain adaptation (UDA) studies the problem of transferring models trained on one labeled source domain to another unlabeled target domain. In this article, we focus on UDA in visual emotion analysis for both emotion distribution learning and dominant emotion classification. Specifically, we design a novel end-to-end cycle-consistent adversarial model, called CycleEmotionGAN++. First, we generate an adapted domain to align the source and target domains on the pixel level by improving CycleGAN with a multiscale structured cycle-consistency loss. During the image translation, we propose a dynamic emotional semantic consistency loss to preserve the emotion labels of the source images. Second, we train a transferable task classifier on the adapted domain with feature-level alignment between the adapted and target domains. We conduct extensive UDA experiments on the Flickr-LDL and Twitter-LDL datasets for distribution learning and ArtPhoto and Flickr and Instagram datasets for emotion classification. The results demonstrate the significant improvements yielded by the proposed CycleEmotionGAN++ compared to state-of-the-art UDA approaches.

关键词： Task analysis visualization Adaptation models Semantics Training Multimedia Web sites Machine learning Unsupervised domain adaptation (UDA) dynamic emotional semantic consistency (DESC) emotion distribution generative adversarial networks (GANs) visual emotion analysis (VEA)

来源：评论

学校读者我要写书评

暂无评论

Relation-Aware Fine-Grained Reasoning Network for Textbook Question Answering

引用

ieee TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2023年第1期34卷 15-27页

作者： Ma, Jie Liu, Jun Wang, Yaxian Li, Junjun Liu, Tongliang Xi An Jiao Tong Univ Sch Comp Sci & Technol Natl Engn Lab Big Data Analyt Xian 710049 Peoples R China Xi An Jiao Tong Univ Sch Comp Sci & Technol Shaanxi Prov Key Lab Satellite & Terr Network Tec Xian 710049 Peoples R China Univ Sydney Sch Comp Sci Fac Engn Trustworthy Machine Learning Lab Darlington NSW 2008 Australia

Textbook question answering (TQA) is a task that one should answer non-diagram and diagram questions accurately, given a large context which consists of abundant diagrams and essays. Although lots of studies have made significant progress in the natural image question answering (QA), they are not applicable to comprehending diagrams and reasoning over the long multimodal context. To address the above issues, we propose a relation-aware fine-grained reasoning (RAFR) network that performs fine-grained reasoning over the nodes of relation-based diagram graphs. Our method uses semantic dependencies and relative positions between nodes in the diagram to construct relation graphs and applies graph attention networks to learn diagram representations. To extract and reason over the multimodal knowledge, we first extract the text that is the most relevant to questions, options, and the instructional diagram which is the most relevant to question diagrams at the word-sentence level and the node-diagram level, respectively. Then, we apply instructional-diagram-guided attention and question-guided attention to reason over the node of question diagrams, respectively. The experimental results show that our proposed method achieves the best performance on the TQA dataset compared with baselines. We also conduct extensive ablation studies to comprehensively analyze the proposed method.

关键词： Semantics Cognition Knowledge discovery Task analysis visualization Shape Frontal lobe Diagram understanding machine reading comprehension (MRC) multi-modality textbook question answering (TQA)

来源：评论

学校读者我要写书评

暂无评论

Heterogeneous Face Interpretable Disentangled Representation for Joint Face Recognition and Synthesis

引用

ieee TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2022年第10期33卷 5611-5625页

作者： Liu, Decheng Gao, Xinbo Peng, Chunlei Wang, Nannan Li, Jie Xidian Univ Sch Elect Engn State Key Lab Integrated Serv Networks Xian 710071 Peoples R China Chongqing Univ Posts & Telecommun Chongqing Key Lab Image Cognit Chongqing 400065 Peoples R China Xidian Univ Sch Cyber Engn State Key Lab Integrated Serv Networks Xian 710071 Peoples R China Xidian Univ Sch Telecommun Engn State Key Lab Integrated Serv Networks Xian 710071 Peoples R China

Heterogeneous faces are acquired with different sensors, which are closer to real-world scenarios and play an important role in the biometric security field. However, heterogeneous face analysis is still a challenging problem due to the large discrepancy between different modalities. Recent works either focus on designing a novel loss function or network architecture to directly extract modality-invariant features or synthesizing the same modality faces initially to decrease the modality gap. Yet, the former always lacks explicit interpretability, and the latter strategy inherently brings in synthesis bias. In this article, we explore to learn the plain interpretable representation for complex heterogeneous faces and simultaneously perform face recognition and synthesis tasks. We propose the heterogeneous face interpretable disentangled representation (HFIDR) that could explicitly interpret dimensions of face representation rather than simple mapping. Benefited from the interpretable structure, we further could extract latent identity information for cross-modality recognition and convert the modality factor to synthesize cross-modality faces. Moreover, we propose a multimodality heterogeneous face interpretable disentangled representation (M-HFIDR) to extend the basic approach suitable for the multimodality face recognition and synthesis. To evaluate the ability of generalization, we construct a novel large-scale face sketch data set. Experimental results on multiple heterogeneous face databases demonstrate the effectiveness of the proposed method.

关键词： Face recognition Task analysis Feature extraction Image recognition Sensors Semantics visualization Disentanglement heterogeneous face interpretable representation joint recognition and synthesis

来源：评论

学校读者我要写书评

暂无评论

High Performance visualization for Astrophysics and Cosmology 33

High Performance Visualization for Astrophysics and Cosmolog...

引用

33rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2025

作者： Tuccari, Nicola Sciacca, Eva Vitello, Fabio Colonnelli, Iacopo Becerra, Yolanda Cintero, Enric Sosa Marin, Guillermo Jaros, Milan Riha, Lubomir Strakos, Petr Trujillo-Gomez, Sebastian Tramontana, Emiliano Wissing, Robert Istituto Nazionale di Astrofisica Catania Italy Università di Torino Torino Italy Barcelona Supercomputing Center Barcelona Spain IT4Innovations National Supercomputing Center Ostrava Czech Republic Heidelberg Institute for Theoretical Studies Heidelberg Germany Università degli Studi di Catania Italy University of Oslo Institute of Theoretical Astrophysics Oslo Norway

ISBN: (纸本)9798331524937

Modern Astrophysics and Cosmology (A&C) projects produce immense data volumes, necessitating advanced software tools for data access, storage, and analysis. visualization Interface for the Virtual Observatory (VisIVO) is one such tool enabling multi-dimensional data analysis and knowledge discovery across complex astrophysical datasets. Leveraging containerization and virtualization, VisIVO has been deployed on various distributed computing platforms. Additionally, Blender, an open-source 3D suite, provides robust tools for rendering and processing volumetric data, making it suitable for visualizing complex datasets. At the SPACE Center of Excellence these tools are being adapted for high-performance visualization of cosmological simulations performed with GADGET and ChaNGa on pre-exascale systems. However, implementing high-performance visualization on diverse HPC platforms presents several challenges, including hardware and software compatibility, data management, scalability, performance portability, and efficient resource allocation. This paper outlines strategies to integrate VisIVO with workflow frameworks and streaming platforms to address these challenges. Workflow frameworks enhance portability, scheduling, and reproducibility of visualization workflows on pre-exascale systems used in A&C simulations. We also discuss the use of streaming platforms to enable concurrent (i.e. in-situ) analysis and visualization of simulations, reducing the need to store full simulation data by leveraging distributed databases that stream the output data in real time. Lastly, we present an adaptation of Blender to handle large-scale particle-based astrophysical data, offering high-quality visualization with interactive exploration capabilities. © 2025 ieee.

关键词： Dark Matter

来源：评论

学校读者我要写书评

暂无评论

Cytosplore EvoViewer: Visual Analytics of Conserved Evolutionary Patterns in multi-species single-cell sequencing data

Cytosplore EvoViewer: Visual Analytics of Conserved Evolutio...

引用

Pacific (formerly Asia-Pacific APVIS) visualization symposium

作者： Soumyadeep Basu Morgan Wirthlin Jeroen Eggermont Thomas Kroes Boudewijn Lelieveldt Ed Lein Trygve Bakken Thomas Höllt Leiden University Medical Center Delft University of Technology Allen Institute for Brain Science

ISBN: (数字)9798331505813

ISBN: (纸本)9798331505820

Single-cell transcriptomics has enhanced our understanding of the brain’s cellular composition. Biologists now analyze complex datasets to explore how marker genes influence biological processes, genetic variations, and phenotypic traits. A challenge is comparing these datasets across species to detect subtle differences or similarities to evolutionary development. Here, we present Cytosplore EvoViewer to facilitate examining relationships between transcriptomic datasets across species, simplifying the analysis of marker gene regulation and its impact on biological functions and integrating these findings with prior evolutionary knowledge or species-specific traits. We conducted a design study, including domain analysis, implementation of the results into Cytosplore EvoViewer, and an expert evaluation. Cytosplore EvoViewer offers valuable insights into genetic variations and evolutionary dynamics, helping to understand the diversity and the unity within diversity across species and their evolutionary *** Cytosplore EvoViewer installer application can be downloaded from the Cytosplore Viewer website 1 , and its source code is available on the ManiVault Studio GitHub 2 . 1 https://*** 2 https://***/ManiVaultStudio/CytosploreEvoViewer

关键词： Sequential analysis High dimensional data Visual analytics Source coding Transcriptomics data visualization Collaboration Genetics Regulation Biological information theory

来源：评论

学校读者我要写书评

暂无评论

Explorative Visual analysis of Spatio-temporal Regions to Detect Hemodynamic Biomarker Candidates

Explorative Visual Analysis of Spatio-temporal Regions to De...

引用

24th Eurographics Conference on visualization, EuroVis 2022

作者： Derstroff, Adrian Leistikow, Simon Nahardani, Ali Ebrahimi, Mahyasadat Hoerr, Verena Linsen, Lars University of Münster Germany Medical Physics Group Institute of Diagnostic and Interventional Radiology Jena University Hospital Friedrich Schiller University Jena Jena Germany Heart Center Bonn Department of Internal Medicine II University Hospital Bonn Bonn Germany Clinic for Radiology University Hospital Münster Münster Germany Leipniz-Institute of Photonic Technology Jena Germany Friedrich Schiller University Jena Germany Institute of Physical Chemistry Abbe Center of Photonics Jena Germany

ISBN: (纸本)9783038681854

Biomarkers are measurable biological properties that allow for distinguishing subjects of different cohorts such as healthy vs. diseased. In the context of diagnosing diseases of the cardiovascular system, researchers aim - among others - at detecting biomarkers in the form of spatio-temporal regions of blood flow obtained by medical imaging or of derived hemodynamical parameters. As the search space for such biomarkers in time-varying volumetric multi-field data is extremely large, we present an interactive visual exploration system to support the analysis of the potential of spatio-temporal regions to discriminate cohorts. © 2022 The Author(s) Eurographics Proceedings © 2022 The Eurographics Association.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Exploratory data analysis in Electronic Health Records Graphs: Intuitive Features and visualization Tools

Exploratory Data Analysis in Electronic Health Records Graph...

引用

Annual ieee symposium on Computer-Based Medical Systems

作者： Mirela T. Cazzolato Marco Antonio Gutierrez Caetano Traina Christos Faloutsos Agma J.M. Traina Institute of Mathematics and Computer Science - University of São Paulo (ICMC-USP) São Carlos Brazil The Heart Institute (InCor) - Clinical Hospital of Faculty of Medicine University of São Paulo (HC-FMUSP) São Paulo Brazil Carnegie Mellon University (CMU) Pittsburgh USA

Given a large, unlabeled set of Electronic Health Records (EHRs) acquired from multiple hospitals, how can we analyze the available entities and identify relationships in the data? Also, how can we perform Exploratory data analysis (EDA) over such EHR data? Many medical institutions generate EHRs as tabular data with entities and attributes in common. However, due to a large number of records, attributes, and high cardinality, exploring the different datasets and finding patterns and insights become laborious and prone to errors. In this work, we propose GraF- Eda for EDA over EHR data from different institutions. GraF-EDA models EHRs as time-evolving graphs, allowing the interoperability of such data into a single representation. We extract meaningful features from the graph nodes and provide intuitive visualizations to improve data explainability. We evaluate GraF-EDA with four COVID-19 datasets from hospitals of the São Paulo state, Brazil, resulting in million-scale graphs. Our method identified correlations, similarities and dissimilarities among medical treatments, exams, clinics, and outcomes. With the visual tools provided by GraF-EDA, we were able to spot cases of interest and check more details about them. Our results indicate that GraF-EDA is a fast, effective, open-sourced tool for EDA of EHRs from multiple institutions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Discriminative Cross-Aligned Variational Autoencoder for Zero-Shot Learning

引用

ieee TRANSACTIONS ON CYBERNETICS 2023年第6期53卷 3794-3805页

作者： Liu, Yang Gao, Xinbo Han, Jungong Shao, Ling Xidian Univ State Key Lab Integrated Serv Networks Xian 710071 Peoples R China Shanghai Jiao Tong Univ Key Lab Artificial Intelligence Minist Educ Shanghai 200240 Peoples R China Shandong Univ Key Lab Cryptol Technol & Informat Secur Minist Educ Jinan 250100 Peoples R China Southeast Univ Key Lab Comp Network & Informat Integrat Minist Educ Nanjing 211189 Peoples R China Chongqing Univ Posts & Telecommun Chongqing Key Lab Image Cognit Chongqing 400065 Peoples R China Aberystwyth Univ Dept Comp Sci Aberystwyth SY23 3DB Dyfed Wales Incept Inst Artificial Intelligence Abu Dhabi U Arab Emirates

Zero-shot learning (ZSL) aims to classify unseen samples based on the relationship between the learned visual features and semantic features. Traditional ZSL methods typically capture the underlying multimodal data structures by learning an embedding function between the visual space and the semantic space with the Euclidean metric. However, these models suffer from the hubness problem and domain bias problem, which leads to unsatisfactory performance, especially in the generalized ZSL (GZSL) task. To tackle such a problem, we formulate a discriminative cross-aligned variational autoencoder (DCA-VAE) for ZSL. The proposed model effectively utilizes a modified cross-modal-alignment variational autoencoder (VAE) to transform both visual features and semantic features obtained by the discriminative cosine metric into latent features. The key to our method is that we collect principal discriminative information from visual and semantic features to construct latent features which contain the discriminative multimodal information associated with unseen samples. Finally, the proposed model DCA-VAE is validated on six benchmarks including the large dataset ImageNet, and several experimental results demonstrate the superiority of DCA-VAE over most existing embedding or generative ZSL models on the standard ZSL and the more realistic GZSL tasks.

关键词： visualization Semantics Task analysis Training Solid modeling Prototypes Generative adversarial networks Autoencoder cosine discriminative latent zero-shot learning (ZSL)

来源：评论

学校读者我要写书评

暂无评论

Reimagining TaxiVis through an Immersive Space-Time Cube metaphor and reflecting on potential benefits of Immersive Analytics for urban data exploration

Reimagining TaxiVis through an Immersive Space-Time Cube met...

引用

ieee Annual International symposium Virtual Reality

作者： Jorge Wagner Claudio T. Silva Wolfgang Stuerzlinger Luciana Nedel Federal University of Rio Grande do Sul New York University Simon Fraser University

Current visualization research has identified the potential of more immersive settings for data exploration, leveraging VR and AR technologies. To explore how a traditional visualization system could be adapted into an immersive framework, and how it could benefit from this, we decided to revisit a landmark paper presented ten years ago at ieee VIS. TaxiVis, by Ferreira et al., enabled interactive spatio-temporal querying of a large dataset of taxi trips in New York City. Here, we reimagine how TaxiVis’ functionalities could be implemented and extended in a 3D immersive environment. Among the unique features we identify as being enabled by the Immersive TaxiVis prototype are alternative uses of the additional visual dimension, a fully visual 3D spatio-temporal query framework, and the opportunity to explore the data at different scales and frames of reference. By revisiting the case studies from the original paper, we demonstrate workflows that can benefit from this immersive perspective. Through reporting on our experience, and on the vision and reasoning behind our design decisions, we hope to contribute to the debate on how conventional and immersive visualization paradigms can complement each other and on how the exploration of urban datasets can be facilitated in the coming years.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Usability Study on data Provenance visualization Approaches 15

A Usability Study on Data Provenance Visualization Approache...

引用

15th Turkish National Software Engineering symposium (UYMS)

作者： Yazici, Ilkay Melek Aktas, Mehmet S. Yildiz Tech Univ Comp Engn Dept Istanbul Turkey

ISBN: (纸本)9781665410700

data visualization becomes an arising need to the end-users to assess, analyze, and reveal the data's insights, especially in the Big data age. The amount of data generated by platforms and systems over the world increases rapidly. As an illustration, in Social Media and e-Science, domain data increases at a tremendous amount. data provenance is a sort of metadata on the origin and life cycle of data is the issue of both domains for data analysis. Therefore, this requires further researches to ensure better understanding and efficient data interpretation. Our former comprehensive study has addressed several visualization approaches to meet these needs, such as data summarization and graph comparison in the e-Science domain. Proposed approaches implemented into a visualization tool to illustrate their applicability for the end-users. One of the objections was to involve end-users in the visualization process as much as possible. In this study, we present a detailed user study to assess our implemented visualization environment concerning Human Machine Interface (HMI) from the end-user perspective. Implemented visualization approaches are evaluated using task-level satisfaction and general-level transaction user surveys with well-defined study cases. In addition, System Usability Scale metrics are applied to measure the SUS score of the overall visualization system. The user study experiments proof that the proposed visualization approaches are applicable, and processing overhead is significantly low by facilitating the visualization system.

关键词： data provenance data visualization provenance visualization Human Machine Interface user study

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：