检索结果-内蒙古大学图书馆

25th IEEE International Conference on Computer Supported Cooperative Work in Design, CSCWD 2022

作者： Lin, Kunda Xu, Xiaolong Jiang, Yu School of Computer Science Nanjing University of Posts and Telecommunications Jiangsu Key Laboratory of Big Data Security and Intelligent Processing Nanjing China

ISBN: (纸本)9781665405270

Encrypted network traffic classification is an essential task in modern communications, which is used in a wide range of applications, such as network resource allocation, QoS (Quality of Service), malicious detection, etc. With the continuous evolution of network technology, approaches used to classify network encrypted traffic have become increasingly complex, and model training relies heavily on a large amount of labeled data. However, the acquisition of correct and massive labeled data of network traffic in the real environment remains a major challenge in this field. On the opposite, unlabeled traffic is extremely easy to obtain in a network. Therefore, the effective use of unlabeled data is of great significance to the development of modern communications. In this paper, we propose the Sauce model, which can effectively use unlabeled information to obtain high-quality clustering space. It is composed of Aux (auxiliary network) and AE (auto encoder), where Aux is designed for collaborative training with AE, affecting the distribution of samples in the latent space generated by AE. Sauce uses t-SNE (t- distributed stochastic neighbor embedding) to perform secondary dimensionality reduction on the latent code, and then uses a clustering algorithm for cluster analysis. Besides, sauce applies the self-training technique, replacing the real labeled data with pseudo-labeled data to reduce the reliance on labeled data and improve the utilization of unannotated data. We conduct experiments on two real network datasets. The experiments show that Sauce can achieve clustering accuracies of 98.4% (VPN dataset) and 98.0% (TOR dataset), outperforming other unsupervised or semi-supervised learning methods. © 2022 IEEE.

关键词： Cluster analysis

来源：评论

学校读者我要写书评

暂无评论

KAN-Face: Efficient Resource Usage and Precision Lip-Sync in Talking Head Generation

KAN-Face: Efficient Resource Usage and Precision Lip-Sync in...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Feng, Guanwen Jin, Siyu Qian, Zhihao Li, Yunan Miao, Qiguang School of Computer Science and Technology Xidian University Shaanxi Xi'an710071 China Xi'an Key Laboratory of Big Data and Intelligent Vision Xidian University Shaanxi Xi'an710071 China Key Laboratory of Collaborative Intelligence Systems Ministry of Education Xidian University Xi'an710071 China

ISBN: (纸本)9798350368741

Despite significant progress in NeRF-based talking head generation, problems like poor lip synchronization and inefficient resource usage remain. To solve these, we propose KAN-Face, a lightweight framework. In preprocessing, we introduce a Lip-Sync Enhancement Module that uses Wav2Lip to extract high-resolution audio features and map them to an explicit intermediate representation, ensuring precise lip movement alignment with the speaker's identity. These predicted lip-sync features are combined with fundamental audio-extracted lip features and injected into the rendering module to improve synchronization. For rendering, we introduce FastKAN to map spatial points to color values. As a variant of KAN, FastKAN's sensitivity to 3D scenes and efficient structure enable precise, fast color prediction. Our framework reduces resource consumption while enhancing lip-sync accuracy and facial reconstruction, making it ideal for talking head generation tasks in resource-limited settings. Project: https://***/KAN-Face/. © 2025 IEEE.

关键词： Face Animation KAN Network Low Resource Neural Radiance Fields Talking Head Generation

来源：评论

学校读者我要写书评

暂无评论

Text-Image Scene Graph Fusion for Multimodal Named Entity Recognition

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第6期5卷 2828-2839页

作者： Cheng, Jian Long, Kaifang Zhang, Shuang Zhang, Tian Ma, Lianbo Cheng, Shi Guo, Yinan State Key Laboratory for Intelligent Coal Mining and Strata Control Research Institute of Mine Big Data Chinese Institute of Coal Science Beijing100013 China Northeastern University College of Software Shenyang110819 China Shaanxi Normal University College of Computer Science Xi'an710062 China China University of Mining and Technology Xuzhou221116 China

With the popularity and widespread use of social media platforms, such as Twitter and Facebook, massive amounts of text and image information posted by a variety of users have flooded these social media platforms. Thus, multimodal named entity recognition (MNER) has become a research hotspot for the task of extracting named entities from multimodal data. Empirically, the visual clues unrelated to text data may introduce uncertain or even negative impacts on the named entity recognition. The considerations of the relevance of multimodal data have been ignored in the previous studies. In this article, to effectively measure the relationship between text data and visual cues for improving the accuracy of named entities, we propose a text-image scene graph fusion (TISGF) approach with a text-image similarity assessment module (TISA) and a text-image fusion module (TIF) for MNER. Specifically, we first construct two (visual and textual) scene graphs to exploit the joint features of objects and relations in text and image and encode the two scene graphs separately using a specific encoder pair. In this way, we can obtain both object-level and relationship-level cross-modal features. Subsequently, TISA is used to compute the similarity of the image and text data and to determine the proportion of visual information that will be retained for fusion. Finally, we use TIF to achieve a unified multimodal representation of each word and predict the entity type using conditional random fields. Extensive experiment results on two public datasets demonstrate the effectiveness and competitiveness of our proposed method for the MNER task. © 2020 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Carbon emission reduction enabled by informatization construction: an analysis of spatial effects based on China’s experience

引用

Environmental Science and Pollution Research 2024年第24期31卷 35595-35608页

作者： He, Zhuo Ma, Shaojuan Deng, Zixin Meng, Ying College of Business Administration Nanchang Institute of Technology Jiangxi Nanchang330032 China School of Mathematics and Information Science Northern Minzu University Ningxia Yinchuan750021 China Ningxia Key Laboratory of Intelligent Information and Big Data Processing Ningxia Yinchuan750021 China School of Statistics Capital University of Economics and Business Beijing100071 China

The "dual-carbon" objective presents a huge challenge for China and the world, with profound implications for the advancement of China’s eco-friendly economy. Additionally, informatization development has a significant impact on the level of carbon emissions in both local and neighbouring regions. Therefore, we employ panel data from 30 provinces in China spanning the years 2012 to 2021, and use the Kernel density estimate and Moran’s index to explore informatization level and carbon emissions space agglomeration characteristics. We elucidate the nonlinear relationship and heterogeneity between informatization improvement and carbon emissions based on the spatial Durbin model. The primary findings are as follows. Firstly, we discover a distinct spatial clustering phenomenon which the informatization level is high in coastal areas and low in inland areas, whereas carbon emissions are low in the south and high in the north. Secondly, the effect of the informatization level on carbon emissions is shown as a U-shaped and non-linear correlation, signifying inhibitory and subsequently promoting phases. Thirdly, we reveal the negative influence on carbon emissions caused by spatial lag terms of the informatization level, and find that a higher local informatization level will have an inhibitory effect on carbon emissions in neighbouring areas. Finally, there is a spatial heterogeneity in the impact of the informatization level on carbon emissions, which presents the U-shaped relation between informatization level and carbon emissions varies across the North-South subregion and the three major economic subregion of China. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

关键词： Informatization

来源：评论

学校读者我要写书评

暂无评论

Deep contrastive learning based hybrid network for Typhoon intensity classification

引用

EXPERT SYSTEMS WITH APPLICATIONS 2024年第PartB期252卷

作者： Yin, Pengshuai Fang, Yupeng Chen, Huanxin Huang, Huichou Wan, Qilin Wu, Qingyao Guangdong Lab Artificial Intelligence & Digital Ec Shenzhen 518123 Peoples R China South China Univ Technol Sch Software Engn Guangzhou 510006 Peoples R China City Univ HongKong Hong Kong Peoples R China CMA Guangzhou Inst Trop & Marine Meteorol Guangzhou 510641 Peoples R China Minist Educ Key Lab Big Data & Intelligent Robot Guangzhou 510006 Peoples R China

The accurate classification of Typhoon Intensity (TI) based on cloud patterns in satellite images is crucial for effective disaster warning and management. However, this task is particularly challenging due to the striking visual similarities among different sub -classes of typhoons, coupled with a long-tailed distribution of class occurrences. Existing deep learning methods often struggle with biased classification stemming from imbalanced datasets and face difficulty discerning subtle differences between categories. This paper proposes a novel solution to these challenges through a hybrid framework that integrates contrastive learning and classifier learning. Contrastive learning is employed to increase the separation of similar sub -classes within the feature space, while classifier learning aims to train a discriminative and unbiased classifier. The proposed approach is evaluated extensively on the publicly available DeepTI dataset, demonstrating enhanced performance for both prominent and less frequent classes. The model achieves the 70.88% accuracy for seven categories typhoon intensity classification. The code implementation is at https://***/chen-huanxin/contrastive-hybridnetwork.

关键词： Typhoon intensity Image classification Deep learning

来源：评论

学校读者我要写书评

暂无评论

Robust Latent Factor Analysis for Precise Representation of High-Dimensional and Sparse data

引用

IEEE/CAA Journal of Automatica Sinica 2021年第4期8卷 796-805页

作者： Di Wu Xin Luo the Chongqing Key Laboratory of Big Data and Intelligent Computing Chongqing Institute of Green and Intelligent TechnologyChinese Academy of SciencesChongqing 400714 the Chongqing School University of Chinese Academy of SciencesChongqing 400714China

High-dimensional and sparse(HiDS)matrices commonly arise in various industrial applications,e.g.,recommender systems(RSs),social networks,and wireless sensor *** they contain rich information,how to accurately represent them is of great significance.A latent factor(LF)model is one of the most popular and successful ways to address this *** LF models mostly adopt L2-norm-oriented Loss to represent an HiDS matrix,i.e.,they sum the errors between observed data and predicted ones with *** L2-norm is sensitive to outlier ***,outlier data usually exist in such *** example,an HiDS matrix from RSs commonly contains many outlier ratings due to some heedless/malicious *** address this issue,this work proposes a smooth L1-norm-oriented latent factor(SL-LF)*** main idea is to adopt smooth L1-norm rather than L2-norm to form its Loss,making it have both strong robustness and high accuracy in predicting the missing data of an HiDS *** results on eight HiDS matrices generated by industrial applications verify that the proposed SL-LF model not only is robust to the outlier data but also has significantly higher prediction accuracy than state-of-the-art models when they are used to predict the missing data of HiDS matrices.

关键词： High-dimensional and sparse matrix L1-norm L2 norm latent factor model recommender system smooth L1-norm

来源：评论

学校读者我要写书评

暂无评论

G-NeRF: Geometry-enhanced Novel View Synthesis from Single-View Images

G-NeRF: Geometry-enhanced Novel View Synthesis from Single-V...

引用

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Zixiong Huang Qi Chen Libo Sun Yifan Yang Naizhou Wang Qi Wu Mingkui Tan South China University of Technology Guangzhou Shiyuan Electronics Co. Ltd The University of Adelaide Pazhou Lab Key Laboratory of Big Data and Intelligent Robot Ministry of Education

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Novel view synthesis aims to generate new view images of a given view image collection. Recent attempts address this problem relying on 3D geometry priors (e.g., shapes, sizes, and positions) learned from multi-view images. However, such methods encounter the following limitations: 1) they require a set of multi-view images as training data for a specific scene (e.g., face, car or chair), which is often un-available in many real-world scenarios; 2) they fail to ex-tract the geometry priors from single-view images due to the lack of multi-view supervision. In this paper, we pro-pose a Geometry-enhanced NeRF (G-NeRF), which seeks to enhance the geometry priors by a geometry-guided multi-view synthesis approach, followed by a depth-aware training. In the synthesis process, inspired that existing 3D GAN models can unconditionally synthesize high-fidelity multi-view images, we seek to adopt off-the-shelf 3D GAN models, such as EG3D, as a free source to provide geometry priors through synthesizing multi-view data. Simultaneously, to further improve the geometry quality of the synthetic data, we introduce a truncation method to effectively sample la-tent codes within 3D GAN models. To tackle the absence of multi-view supervision for single-view images, we design the depth-aware training approach, incorporating a depth-aware discriminator to guide geometry priors through depth maps. Experiments demonstrate the effectiveness of our method in terms of both qualitative and quantitative results.

关键词： Geometry Training Solid modeling Three-dimensional displays Shape Training data Neural radiance field

来源：评论

学校读者我要写书评

暂无评论

G-NeRF: Geometry-enhanced Novel View Synthesis from Single-View Images

G-NeRF: Geometry-enhanced Novel View Synthesis from Single-V...

引用

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Huang, Zixiong Chen, Qi Sung, Libo Yang, Yifan Wang, Naizhou Wu, Qi Tan, Mingkui South China Univ Technol Guangzhou Peoples R China Univ Adelaide Adelaide SA Australia Guangzhou Shiyuan Elect Co Ltd Guangzhou Peoples R China Pazhou Lab Pazhou Peoples R China Minist Educ Key Lab Big Data & Intelligent Robot Beijing Peoples R China

ISBN: (纸本)9798350353006

Novel view synthesis aims to generate new view images of a given view image collection. Recent attempts address this problem relying on 3D geometry priors (e.g., shapes, sizes, and positions) learned from multi-view images. However, such methods encounter the following limitations: 1) they require a set of multi-view images as training data for a specific scene (e.g., face, car or chair), which is often unavailable in many real-world scenarios;2) they fail to ex-tract the geometry priors from single-view images due to the lack of multi-view supervision. In this paper, we propose a Geometry-enhanced NeRF (G-NeRF), which seeks to enhance the geometry priors by a geometry-guided multi-view synthesis approach, followed by a depth-aware train-ing. In the synthesis process, inspired that existing 3D GAN models can unconditionally synthesize high-fidelity multi-view images, we seek to adopt off-the-shelf 3D GAN models, such as EG3D, as a free source to provide geometry priors through synthesizing multi-view data. Simultaneously, to further improve the geometry quality of the synthetic data, we introduce a truncation method to effectively sample latent codes within 3D GAN models. To tackle the absence of multi-view supervision for single-view images, we design the depth-aware training approach, incorporating a depth-aware discriminator to guide geometry priors through depth maps. Experiments demonstrate the effectiveness of our method in terms of both qualitative and quantitative results.

关键词： NeRF Single-view Reconstruction

来源：评论

学校读者我要写书评

暂无评论

Enhancing Perception Capabilities of Multimodal LLMs with Training-Free Fusion

arXiv

引用

arXiv 2024年

作者： Chen, Zhuokun Hu, Jinwu Deng, Zeshuai Wang, Yufeng Zhuang, Bohan Tan, Mingkui South China University of Technology China Pazhou Lab China ZIP Lab Zhejiang University China Key Laboratory of Big Data and Intelligent Robot Ministry of Education China

Multimodal LLMs (MLLMs) equip language models with visual capabilities by aligning vision encoders with language models. Existing methods to enhance the visual perception of MLLMs often involve designing more powerful vision encoders, which requires exploring a vast design space and re-aligning each potential encoder with the language model, resulting in prohibitively high training costs. In this paper, we introduce VisionFuse, a novel integration framework that efficiently utilizes multiple vision encoders from off-the-shelf MLLMs to enhance visual perception without requiring additional training. Our approach is motivated by the observation that different MLLMs tend to focus on distinct regions given the same query and image. Moreover, we find that the feature distributions of vision encoders within an MLLM family, a group of MLLMs sharing the same pretrained LLM, are highly aligned. Building on these insights, VisionFuse enriches the visual context by concatenating the tokens generated by the vision encoders of selected MLLMs within a family. By merging the parameters of language models from these MLLMs, VisionFuse allows a single language model to align with various vision encoders, significantly reducing deployment overhead. We conduct comprehensive evaluations across multiple multimodal benchmarks using various MLLM combinations, demonstrating substantial improvements in multimodal tasks. Notably, when integrating MiniGemini-8B and SLIME-8B, VisionFuse achieves an average performance increase of over 4%. © 2024, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Obstacle-Avoidance Path Planning of robot Arm Based on Improved RRT Algorithm 8

Obstacle-Avoidance Path Planning of Robot Arm Based on Impro...

引用

8th Asia-Pacific Conference on intelligent robot Systems, ACIRS 2023

作者： Hu, Chong Mu, Chunyang Xing, Ma Zhang, Chuntao Zhou, Wenya Yang, Ke School of Electrical and Information Engineering North Minzu University Yinchuan China College of Mechanical and Electrical Engineering North Minzu University The Key Laboratory of Intelligent Information and Big Data Processing of Ningxia Province Yinchuan China School of Electrical and Information Engineering North Minzu University The Key Laboratory of Intelligent Information and Big Data Processing of Ningxia Province Yinchuan China College of Mechanical and Electrical Engineering North Minzu University Yinchuan China

ISBN: (纸本)9798350302851

In the application of RRT (Rapidly-exploring Random Trees) algorithm in obstacle avoidance path planning of redundant robot arms in high-dimensional space, the sampling area of random sampling points is large, the search time is long, the path twists and turns, and the redundant points are too many. In this paper, an efficient sampling RRT algorithm with adaptive step size and fast automatic convergence is proposed. Firstly, by adding the optimized gravity function, the step size of the extended tree is changed adaptively and the convergence is stronger. Secondly, the limited sampling of parent node area expansion is proposed to avoid useless point sampling and repeated point sampling, so that the utilization of sampling points is greatly improved. Finally, by changing the way of parent node reconnection and re-selection, the path tortuous degree is lower, the path planning cost and redundancy points are reduced, and MATLAB software is used to carry out the path planning simulation experiment. By comparing RRT, gravitational field RTT and RRT∗ algorithm, the improved RRT algorithm has significantly improved in distance, time and number of nodes. It has important engineering application value. © 2023 IEEE.

关键词： Motion planning

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：