检索结果-内蒙古大学图书馆

S-Faster R-CNN: Intra-Spectral Similarity Learning for Audio Copy-Move Forgery Localization in IoT Security

IEEE Internet of Things Journal 2025年

作者： Qiu, Xin Shi, Canghong Li, Xiaojie Wu, Min Abdullahi, Sani M. Liu, Yong Xihua University School of Computer and Software Engineering Chengdu610039 China Chengdu University of Information Technology School of Computer Science Chengdu610225 China University of Southern Denmark M Rsk McKinney Moller Institute Odense5230 Denmark Shanghai University School of Communication and Information Systems 200444 China

In the Internet of Audio Things, communication security of the audio control terminal is vulnerable to copy-move threats, and detecting and locating audio copy-move forgery remains challenging nowadays. The forgery detection method based on deep learning achieves higher detection accuracy but fails to localize forged regions. To address this issue, this article proposes an S-Faster R-CNN model for audio copy-move forgery detection and localization. We integrate a novel Similarity Computation Module (SCM) into the Faster R-CNN framework, forming the S-Faster R-CNN model. Obtaining the integration of the SCM, which allows the S-Faster R-CNN to precisely localize forgery regions within the spectrogram. Finally, the image coordinate transformation algorithm is used to map these forged regions to the corresponding locations of the original audio waveform, thus completing the audio copy-move forgery detection and localization. Evaluated on three datasets, our method achieves an average recall of 90%, an average precision of 84%, and an average F1-score of 87%, respectively. Experimental results indicate that the S-Faster R-CNN outperforms state-of-the-art methods in both forgery detection accuracy and especially in localization. Moreover, the proposed method shows good robustness under multiple post-processing. © 2014 IEEE.

关键词： Spectrographs

来源：评论

学校读者我要写书评

暂无评论

InsTaG: Learning Personalized 3D Talking Head from Few-Second Video

arXiv

引用

arXiv 2025年

作者： Li, Jiahe Zhang, Jiawei Bai, Xiao Zheng, Jin Zhou, Jun Gu, Lin School of Computer Science and Engineering State Key Laboratory of Complex & Critical Software Environment Jiangxi Research Institute Beihang University China School of Information and Communication Technology Griffith University Australia RIKEN AIP Japan The University of Tokyo Japan

Despite exhibiting impressive performance in synthesizing lifelike personalized 3D talking heads, prevailing methods based on radiance fields suffer from high demands for training data and time for each new identity. This paper introduces InsTaG, a 3D talking head synthesis framework that allows a fast learning of realistic personalized 3D talking head from few training data. Built upon a lightweight 3DGS person-specific synthesizer with universal motion priors, InsTaG achieves high-quality and fast adaptation while preserving high-level personalization and efficiency. As preparation, we first propose an Identity-Free Pre-training strategy that enables the pre-training of the person-specific model and encourages the collection of universal motion priors from long-video data corpus. To fully exploit the universal motion priors to learn an unseen new identity, we then present a Motion-Aligned Adaptation strategy to adaptively align the target head to the pre-trained field, and constrain a robust dynamic head structure under few training data. Experiments demonstrate our outstanding performance and efficiency under various data scenarios to render high-quality personalized talking heads. Project page: https://***/InsTaG/ . Copyright © 2025, The Authors. All rights reserved.

关键词： 3D modeling

来源：评论

学校读者我要写书评

暂无评论

A Correlation And Order-Aware Rule Learning Method For Knowledge Graph Reasoning

A Correlation And Order-Aware Rule Learning Method For Knowl...

引用

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Yuefeng He Yang He Xu Zheng Bei Hui Bin Tian Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China Shenzhen School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu School of Information and Software Engineering University of Electronic Science and Technology of China Chengdu Naval University of Engineering Wuhan

Mining high-quality logical rules is crucial as they can provide beneficial interpretability for predictions. Recent methods that incorporate logical rules into learning tasks have been proven to yield high-quality logical rules successfully. However, existing methods either rely on the rule instances observed to support rule mining, or simply embed the rule head and the rule body to learn from them. Additionally, they can not fully utilize the rich semantic information contained in logical rules and overlook the intrinsic correlations between all relations within the domain. In this paper, we propose a model called Correlation and order-Aware Rule Learning (CARL) that captures deeper semantic information in rules by allowing relations to be co-aware of each other and paying attention to logical sequence sensitivity. CARL utilizes semantic consistency between the rule body and rule head as its learning objective, continuously introducing more semantic information and logically simplifying the rule body while considering logical sequence sensitivity. We explored the internal correlations between domain relations and used the thought of knowledge distillation to simplify modules so that relations in CARL can share or perceive each other’s information or state efficiently. Experiments on link prediction tasks have demonstrated that CARL can learn higher-quality rules and yield state-of-the-art results on four popular public datasets. https://***/burning5112/CARL

关键词：

来源：评论

学校读者我要写书评

暂无评论

Intrusion Detection in Smart IoT Devices for People with Disabilities 8

Intrusion Detection in Smart IoT Devices for People with Dis...

引用

8th IEEE International Smart Cities Conference, ISC2 2022

作者： Naveed, Muhammad Usman, Syed Muhammad Satti, Muhammad Islam Aleshaiker, Sama Anwar, Aamir Shaheed Zulfikar Ali Bhutto Institute of Science and Technology Department of Computer Science Islamabad Pakistan Ai Air University Faculty of Computing Department of Creative Technologies Islamabad Pakistan Riphah University Faculty of Computer Science and Software Engineering Islamabad Pakistan School of Computing & Engineering University of West London London United Kingdom

ISBN: (数字)9781665485616

ISBN: (纸本)9781665485616

An intrusion Detection System (IDS) is a system that resides inside the network and monitors all incoming and outgoing traffic. It prevents unethical activities from happening over the network. With the use of IoT devices, network traffic is also increased. Intruders and hackers are attracted to this network because of its low processing power and openness. IoT has transformed diagnostic and monitoring systems for patients in the healthcare industry. However, a secure network is needed for these health care devices. This research proposes a hybrid model to secure the IoT network from external intrusions. The proposed method consists of preprocessing data with the help of normalization and feature selection by removing high correlated features with the help of the Pearson correlation coefficient and Support Vector Machine (SVM) for classification. The proposed approach has achieved an accuracy of 99.3%, precision of 99.1% and an F-1 score of 99.25% on the standard dataset. Results have been compared with state-of-the-art, and the proposed method outperforms all performance measures. © 2022 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

HairDiffusion: Vivid Multi-Colored Hair Editing via Latent Diffusion

arXiv

引用

arXiv 2024年

作者： Zeng, Yu Zhang, Yang Liu, Jiachen Shen, Linlin Deng, Kaijun He, Weizhao Wang, Jinbao Computer Vision Institute School of Computer Science & Software Engineering Shenzhen University China Shenzhen Institute of Artificial Intelligence and Robotics for Society China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Guangdong Provincial Key Laboratory of Intelligent Information Processing China

Hair editing is a critical image synthesis task that aims to edit hair color and hairstyle using text descriptions or reference images, while preserving irrelevant attributes (e.g., identity, background, cloth). Many existing methods are based on StyleGAN to address this task. However, due to the limited spatial distribution of StyleGAN, it struggles with multiple hair color editing and facial preservation. Considering the advancements in diffusion models, we utilize Latent Diffusion Models (LDMs) for hairstyle editing. Our approach introduces Multi-stage Hairstyle Blend (MHB), effectively separating control of hair color and hairstyle in diffusion latent space. Additionally, we train a warping module to align the hair color with the target region. To further enhance multi-color hairstyle editing, we fine-tuned a CLIP model using a multi-color hairstyle dataset. Our method not only tackles the complexity of multi-color hairstyles but also addresses the challenge of preserving original colors during diffusion editing. Extensive experiments showcase the superiority of our method in editing multi-color hairstyles while preserving facial attributes given textual descriptions and reference images. © 2024, CC BY.

关键词： Color image processing

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Rail Surface Defect Detection Method

Unsupervised Rail Surface Defect Detection Method

引用

International Symposium on Environmentally Conscious Design and Inverse Manufacturing (EcoDesign)

作者： ChenYu Zhao Yi Han Ningkang Ouyang YiFan Zhao School of Artificial Intelligence Zhengzhou Software Vocational and Technical College Zhengzhou China Department of Computer Science and Information Anyang Institute of Technology Engineering Anyang China School of Information Engineering North China University of Water Resources and Electric Power Zhengzhou China

ISBN: (数字)9798350363043

ISBN: (纸本)9798350363050

Railroad travel has become an essential part of modern life, therefore, rail surface defect detection has become a problem that cannot be ignored. However, most of the currently existing methods often require a large number of defect samples for learning, which are labor-intensive and very difficult to obtain. In order to alleviate the data dependence of the detection model, this paper proposes an unsupervised method. Specifically, we use Transformer’s attention mechanism to construct the encoder-decoder structure and train it only on normal samples with the purpose of reconstruction, so that those regions that cannot be reconstructed will be regarded as defects at the time of detection. To prevent the model from generalizing its generalization ability to those unknown defective samples, i.e., to avoid overgeneralization, we introduce a noise perturbation in the training so that the model focuses on the reconstruction of normal features. Experiments show that the proposed unsupervised approach outperforms a host of supervised models with low data requirements and can effectively address the current challenges of track defect detection.

关键词： Rails Surface reconstruction Perturbation methods Noise Transformers Data models Decoding Image reconstruction Unsupervised learning Defect detection

来源：评论

学校读者我要写书评

暂无评论

Understanding Older Adults' (Dis)Engagement with Design Materials 25

Understanding Older Adults' (Dis)Engagement with Design Mate...

引用

2025 CHI Conference on Human Factors in Computing Systems, CHI 2025

作者： Pradhan, Alisha Jelen, Ben Thangaraj, Ramprabu Siek, Katie A. Jette, Shannon Lazar, Amanda Department of Informatics Ying Wu College of Computing New Jersey Institute of Technology NewarkNJ United States Computer Science and Software Engineering Rose-Hulman Institute of Technology Terre HauteIN United States Ying Wu College of Computing New Jersey Institute of Technology NewarkNJ United States Luddy School of Informatics Computing and Engineering Indiana University BloomingtonIN United States University of Maryland College ParkMD United States College of Information Studies University of Maryland College ParkMD United States

ISBN: (纸本)9798400713941

Design workshops are a popular approach to include older adults in the technology design process. However, formative design sessions with older adults have had unexpected outcomes such as the non-use of traditional design materials like craft-based prototyping supplies or disengagement from design activities. Analyzing the engagement of 32 older adults across two design workshops, this paper sheds insights on some of these outcomes. Contributing to a growing body of HCI research on understanding older adults' participation in design, we provide an understanding of how design materials can shape older adults' engagement in formative design activities. Our discussion furthers research on understanding who older adults design for and why, argues for a different understanding of creative expression, and offers considerations for choosing design materials. © 2025 Copyright held by the owner/author(s).

关键词： Design for manufacturability

来源：评论

学校读者我要写书评

暂无评论

A Neural Architecture Search Method with Multi-Dimensional Correlation Representation for Fault Diagnosis

A Neural Architecture Search Method with Multi-Dimensional C...

引用

Smart World Congress (SWC), IEEE

作者： Yumeng Liu Tingqi Wang Xu Zheng Ling Tian Hongan Wang Beijing Key Laboratory of Human-Computer Interaction Institute of Software Chinese Academy of Sciences Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China

As deep learning has become more widely used for fault diagnosis, the shortcomings of model transferability and human model design costs are growing increasingly evident. The current work has tackled each of these two issues. Domain adaptation methods can provide transportability by mapping data to a common domain. For the consumption of manually constructed networks, Neural Architecture Search (NAS) is an intelligent solution that automatically select the ideal network structure. Unfortunately, there is little work to connect these two methodologies in fault diagnosis. Furthermore, these approaches are incapable in extracting internal correlations from instrumental signal data. In this paper, we propose a transferrable NAS fault diagnosis method, which maximizes savings in manual costs. We also integrate graph neural networks (GNN) in a stepwise manner into the NAS and domain adaptation to provide a multi-dimension correlation representation. Firstly, we structure the signal data into graph based on temporal relationships. Then, NAS designs automatically for selecting the optimal network structure and useful features. In NAS, graph encoder of GNN is used to establish internal relationships. Finally, domain adaptation provides transferability. At the same time, GNN’s weights in domain adaptation help to further mine data internal associations. We tested our method on two publicly available datasets and performed comparisons. The experimental results show the proposed method has superior performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Generating Targeted Universal Adversarial Perturbation against Automatic Speech Recognition via Phoneme Tailoring

Generating Targeted Universal Adversarial Perturbation again...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Yujun Zhang Yanqu Chen Jiakai Wang Jin Hu Renshuai Tao Xianglong Liu State Key Laboratory of Complex & Critical Software Environment Beihang University China School of Computer Science and Engineering Beihang University China College of Computer Science Beijing University of Technology China Zhongguancun Laboratory China School of Computer and Information Technology Beijing Jiaotong University China Institute of Data Space Hefei Comprehensive National Science Center China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

There is a growing concern about adversarial attacks against automatic speech recognition (ASR) systems. Although research into targeted universal adversarial examples (AEs) has progressed, current methods are constrained by inefficient exploitation of audio features, demonstrating insufficient attack ability and robustness in the physical world. To solve this problem, we propose a phoneme-tailored attack (PTA) to improve the quality of the generated AEs. Specifically, to improve attack ability, we propose a Diverse Audio Composition Enrichment method, which enhances the utilization of audio features through phoneme-level slicing and recombination. To adapt AEs to complex environments, we propose a Natural Noise Pattern Guidance method to align AEs with natural noise patterns to improve their robustness. Experiments show that our method achieves an average accuracy of more than 72.34% and 98% with and without a norm constraint, and also demonstrates excellent performance in terms of generalization across datasets and resilience to MP3 compression.

关键词： Accuracy Target recognition Perturbation methods Noise Robustness Acoustics Speech processing Digital audio players Automatic speech recognition Resilience

来源：评论

学校读者我要写书评

暂无评论

Achieving Efficient and Accurate Segmentation for Autonomous Driving

Achieving Efficient and Accurate Segmentation for Autonomous...

引用

International Symposium on Environmentally Conscious Design and Inverse Manufacturing (EcoDesign)

作者： Ningkang Ouyang Jie Dong Yi Han ChenYu Zhao School of Information Engineering North China University of Water Resources and Electric Power Zhengzhou China Department of Computer Science and Information Anyang Institute of Technology Engineering Anyang China School of Artificial Intelligence Zhengzhou Software Vocational and Technical College Zhengzhou China

ISBN: (数字)9798350363043

ISBN: (纸本)9798350363050

Autonomous driving technology, as a mainstream trend in today’s technological development, holds significant commercial value. Semantic segmentation, a core technology in this field, faces challenges with current mainstream models due to their complex structures, large numbers of parameters, and the need for high-performance hardware. These issues hinder their application in mobile devices. Therefore, this paper proposes a road scene semantic segmentation algorithm based on dynamic asymmetric convolution and capsule networks, named Asymmetric Capsule Convolutional Networks (ACCN). Specifically, the algorithm uses dynamic asymmetric convolution to construct a baseline network that combines dynamic convolution and capsule dynamic routing algorithms to enhance the model’s representational capabilities. Additionally, a dispersed attention module is built to strengthen the dependency between features to achieve more adaptable network weights. Experiments demonstrate that ACCN achieves a good balance between segmentation accuracy and detection speed. On the CamVid and Cityscapes test datasets, the proposed method in this paper achieved an average Intersection over Union (mIoU) of $\mathbf{8 0. 8\%}$ at 55.2 fps, showcasing better trade-offs in segmentation accuracy and efficiency. The experimental results indicate the effectiveness of our method.

关键词： Accuracy Convolution Heuristic algorithms Semantic segmentation Roads Routing Hardware Mobile handsets Vehicle dynamics Autonomous vehicles

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：