检索结果-内蒙古大学图书馆

Modality Unified Attack for Omni-Modality Person Re-Identification

IEEE Transactions on information Forensics and Security 2025年

作者： Bian, Yuan Liu, Min Yi, Yunqi Wang, Xueping Ma, Yunfeng Wang, Yaonan Hunan University National Engineering Research Center of Robot Visual Perception and Control Technology College of Electrical and Information Engineering Hunan Changsha China Hunan Normal University Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing College of Information Science and Engineering Hunan Changsha China

Deep learning based person re-identification (re-id) models have been widely employed in surveillance systems. Recent studies have demonstrated that black-box single-modality and cross-modality re-id models are vulnerable to adversarial examples (AEs), leaving the robustness of multi-modality re-id models unexplored. Due to the lack of knowledge about the specific type of model deployed in the target black-box surveillance system, we aim to generate modality unified AEs for omni-modality (single-, cross- and multi-modality) re-id models. Specifically, we propose a novel Modality Unified Attack method to train modality-specific adversarial generators to generate AEs that effectively attack different omni-modality models. A multi-modality model is adopted as the surrogate model, wherein the features of each modality are perturbed by metric disruption loss before fusion. To collapse the common features of omnimodality models, Cross Modality Simulated Disruption approach is introduced to mimic the cross-modality feature embeddings by intentionally feeding images to non-corresponding modality-specific subnetworks of the surrogate model. Moreover, Multi Modality Collaborative Disruption strategy is devised to facilitate the attacker to comprehensively corrupt the informative content of person images by leveraging a multi modality feature collaborative metric disruption loss. Extensive experiments show that our MUA method can effectively attack the omni-modality re-id models, achieving 55.9%, 24.4%, 49.0% and 62.7% mean mAP Drop Rate, respectively. © 2005-2012 IEEE.

关键词： Modality unified attack Multi Modality Collaborative Disruption Omni-Modality re-id

来源：评论

学校读者我要写书评

暂无评论

Learning to Learn Transferable Generative Attack for Person Re-Identification

引用

IEEE Transactions on Image processing 2025年 PP卷 PP页

作者： Bian, Yuan Liu, Min Wang, Xueping Ma, Yunfeng Wang, Yaonan Hunan University National Engineering Research Center of Robot Visual Perception and Control Technology College of Electrical and Information Engineering Hunan Changsha China Hunan Normal University Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing College of Information Science and Engineering Hunan Changsha China

Deep learning-based person re-identification (reid) models are widely employed in surveillance systems and inevitably inherit the vulnerability of deep networks to adversarial attacks. Existing attacks merely consider cross-dataset and cross-model transferability, ignoring the cross-test capability to perturb models trained in different domains. To powerfully examine the robustness of real-world re-id models, the Meta Transferable Generative Attack (MTGA) method is proposed, which adopts meta-learning optimization to promote the generative attacker producing highly transferable adversarial examples by learning comprehensively simulated transfer-based crossmodel&dataset&test black-box meta attack tasks. Specifically, cross-model&dataset black-box attack tasks are first mimicked by selecting different re-id models and datasets for meta-train and meta-test attack processes. As different models may focus on different feature regions, the Perturbation Random Erasing module is further devised to prevent the attacker from learning to only corrupt model-specific features. To boost the attacker learning to possess cross-test transferability, the Normalization Mix strategy is introduced to imitate diverse feature embedding spaces by mixing multi-domain statistics of target models. Extensive experiments show the superiority of MTGA, especially in cross-model&dataset and cross-model&dataset&test attacks, our MTGA outperforms the SOTA methods by 20.0% and 11.3% on mean mAP drop rate, respectively. © 1992-2012 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

GLFAFormer: DeepFake forgery detection with adaptive feature extract and align

引用

Digital Signal processing: A Review Journal 2025年 165卷

作者： Chen, Long Zhao, Huihuang He, Jiaxin Meng, Weiliang College of Computer Science and Technology Hengyang Normal University Hengyang421002 China Hunan Provincial Key Laboratory of Intelligent Information Processing and Application Hengyang421002 China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China

With the rapid development of generative models, there is an increasing demand for universal fake image detectors. In this paper, we investigate the problem of fake image detection for the synthesis of generative models to detect fake images from multiple generative methods. Recent research methods explore the benefits of pre-trained models and mainly adopt a fixed paradigm of training additional classifiers separately, but we find that the fixed paradigm hinders the full learning of forgery features, leading to insufficient representation learning in the detector. The main reason is that the fixed paradigm pays too much attention to global features and neglects local features, which limits the ability of the model to perceive image details and leads to some information loss or confusion. In this regard, based on the pre-trained visual language space, our method introduces two core designs. First, we designed a Deep Window Triple Attention (DWTA) module. A similar dense sliding window strategy is adopted to capture multi-scale local abnormal patterns, and the sensitivity to generated artifacts is enhanced through the triple attention mechanism. Secondly, we proposed a Cross-Space Feature Alignment (CSFA) module. A two-way interactive channel between global features and local features is established, and the alignment loss function is used to achieve semantic alignment of cross-modal feature spaces. The aligned features are adaptively fused through a gating mechanism to obtain the final adaptive forged features. Experiments demonstrate that our method, when trained solely on ProGAN data, achieves superior cross-generator generalization: it attains an average accuracy of 94.7% on unseen GANs and generalizes to unseen diffusion models with 94% accuracy, surpassing existing methods by 2.1%. The source code is available at https://***/long2580h/GLFAFormer. © 2025 Elsevier Inc.

关键词： Visual languages

来源：评论

学校读者我要写书评

暂无评论

Corona theorem for the Dirichlet-type space

arXiv

引用

arXiv 2023年

作者： Luo, Shuaibing School of Mathematics Hunan Provincial Key Laboratory of Intelligent Information Processing and Applied Mathematics Hunan University Changsha410082 China

This paper utilizes Cauchy’s transform and duality for the Dirichlet-type space D(µ) with positive superharmonic weight Uµ on the unit disk D to establish the corona theorem for the Dirichlet-type multiplier algebra (formula presented) then (formula presented) thereby generalizing Carleson’s corona theorem for M(H2) = H∞ in [7] and Xiao’s corona theorem for M(D) ⊂ H∞ in [25] thanks to (formula presented);Copyright © 2023, The Authors. All rights reserved.

关键词： Algebra

来源：评论

学校读者我要写书评

暂无评论

Geometric optics approximation sampling

arXiv

引用

arXiv 2024年

作者： Sun, Zejun Zheng, Guang-Hui School of Mathematics Hunan University Changsha410082 China School of Mathematics Hunan Provincial Key Laboratory of Intelligent Information Processing and Applied Mathematics Hunan University Changsha410082 China

In this article, we propose a new dimensionality-independent and gradient-free sampler, called Geometric Optics Approximation Sampling, which is based on the reflector antenna system. The core idea is to construct a reflecting surface that redirects rays from a source with a predetermined simpler measure towards a output domain while achieving a desired distribution defined by the projection of a complex target measure of interest. Given a such reflecting surface, one can generate arbitrarily many independent and uncorrelated samples from the target measure simply by dual re-simulating or rays tracing the reflector antenna system and then projecting the traced rays onto target domain. In order to obtain a desired reflecting surface, we use the method of supporting paraboloid to solve the reflector antenna problem that does not require a gradient information regarding the density of the target measure. Furthermore, within the supporting paraboloid method, we utilize a low-discrepancy sequence or a random sequence to discretize the target measure, which in turn yields a dimensionality-independent approach for constructing the reflecting surface. Meanwhile, we present a dual re-simulation or ray tracing method based on its dual reflecting surface, which enables drawing samples from the target measure using the reflector antenna system obtained through the dimensionality-independent method. In theory, we define a geometric optics approximation measure by means of the reflecting surfaces, that is the pushforward of the source measure. We give its well-defined result. In particular, we prove the stability of this measure with respect to the target domain, which in turn ensures the robustness of the numerical sampling. In addition, we establish error bounds between the projection of the numerical geometric optics approximation measure and the target measure under the Wasserstein metrics. Several examples and numerical experiments comparing with measure transport samplers a

关键词： Reflection

来源：评论

学校读者我要写书评

暂无评论

Bayesian inference via geometric optics approximation

arXiv

引用

arXiv 2024年

Markov chain Monte Carlo (MCMC) simulations have been widely used to generate samples from the complex posterior distribution in Bayesian inferences. However, these simulations often require multiple computations of the forward model in the likelihood function for each drawn sample. This computational burden renders MCMC sampling impractical when the forward model is computationally expensive, such as in the case of partial differential equation models. In this paper, we propose a novel sampling approach called the geometric optics approximation method (GOAM) for Bayesian inverse problems, which entirely circumvents the need for MCMC simulations. Our method is rooted in the problem of reflector shape design, which focuses on constructing a reflecting surface that redirects rays from a source, with a predetermined density, towards a target domain while achieving a desired density distribution. The key idea is to consider the unnormalized Bayesian posterior as the density on the target domain within the optical system and define a geometric optics approximation measure with respect to posterior by a reflecting surface. Consequently, once such a reflecting surface is obtained, we can utilize it to draw an arbitrary number of independent and uncorrelated samples from the posterior measure for Bayesian inverse problems. In theory, we have shown that the geometric optics approximation measure is well-posed. The efficiency and robustness of our proposed sampler, employing the geometric optics approximation method, are demonstrated through several numerical examples provided in this paper. Copyright © 2024, The Authors. All rights reserved.

关键词： Numerical methods

来源：评论

学校读者我要写书评

暂无评论

RIESZ TYPE THEOREMS FOR κ-PLURIHARMONIC MAPPINGS, INVARIANT HARMONIC QUASIREGULAR MAPPINGS AND HARMONIC QUASIREGULAR MAPPINGS

arXiv

引用

arXiv 2023年

作者： Chen, Shaolin Huang, Manzi College of Mathematics and Statistics Hengyang Normal University Hunan Hengyang421002 China Hunan Provincial Key Laboratory of Intelligent Information Processing and Application 421002 China Key Laboratory of High Performance Computing and Stochastic Information Processing College of Mathematics and Statistics Hunan Normal University Hunan Changsha410081 China

The main purpose of this paper is to develop some methods to improve and generalize the main results in a recent paper by Liu and Zhu (Adv. Math., 2023, i.e., [19]). The paper consists of two parts. In the first part, we discuss the Riesz type theorem in the setting of n-dimensional complex spaces for all n ≥ 1. In this part, we first introduce the family of κ-pluriharmonic mappings of the n-dimensional complex unit ball. Then we establish two Riesz type theorems for these mappings, which are the n-dimensional versions of Theorems 1.1 and 1.2 in [19], respectively. Furthermore, even when n = 1, our first result shows that the assumption of the real parts of the mappings not being negative (or being negative) in [19, Theorem 1.1] is redundant;and our second result illustrates that the assumption of "quasiconformality" on the mappings in [19, Theorem 1.2] can be replaced by the weaker one of "quasiregularity". In the second part, we investigate the Riesz type theorem in the setting of n-dimensional real spaces for all n ≥ 2. In this part, first, we prove a Riesz type theorem for invariant harmonic quasiregular mappings of the unit n-dimensional real ball. Our result indicates that (i) the range of the parameter p discussed in [19, Theorem 1.3] can be changed from (1, 2) to (1, ∞);(ii) the assumption of the first coordinate functions of the mappings being non-zero in [19, Theorem 1.3] is redundant. In this way, we complete the discussions carried out in [19, Theorems 1.3 and 1.4]. Second, we obtain a Riesz type theorem for harmonic K-quasiregular mappings of the unit n-dimensional real ball. Our result demonstrates that the range of the parameter p discussed in [15, Theorem 2.1] can be changed from (1, 2) to (1, ∞).MSC Codes 30H10, 30C62 Copyright © 2023, The Authors. All rights reserved.

关键词： Harmonic analysis

来源：评论

学校读者我要写书评

暂无评论

Y-Mol: A Multiscale Biomedical Knowledge-Guided Large Language Model for Drug Development

arXiv

引用

arXiv 2024年

作者： Ma, Tengfei Lin, Xuan Li, Tianle Li, Chaoyi Chen, Long Zhou, Peng Cai, Xibao Yang, Xinyu Zeng, Daojian Cao, Dongsheng Zeng, Xiangxiang College of Computer Science and Electronic Engineering Hunan University China Key Laboratory of Intelligent Computing and Information Processing Ministry of Education Xiangtan University China Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing Hunan Normal University China Xiangya School of Pharmaceutical Sciences Central South University China

Large Language Models (LLMs) have recently demonstrated remarkable performance in general tasks across various fields. However, their effectiveness within specific domains such as drug development remains challenges. To solve these challenges, we introduce Y-Mol, forming a well-established LLM paradigm for the flow of drug development. Y-Mol is a multiscale biomedical knowledge-guided LLM designed to accomplish tasks across lead compound discovery, pre-clinic, and clinic prediction. By integrating millions of multiscale biomedical knowledge and using LLaMA2 as the base LLM, Y-Mol augments the reasoning capability in the biomedical domain by learning from a corpus of publications, knowledge graphs, and expert-designed synthetic data. The capability is further enriched with three types of drug-oriented instructions: description-based prompts from processed publications, semantic-based prompts for extracting associations from knowledge graphs, and template-based prompts for understanding expert knowledge from biomedical tools. Besides, Y-Mol offers a set of LLM paradigms that can autonomously execute the downstream tasks across the entire process of drug development, including virtual screening, drug design, pharmacological properties prediction, and drug-related interaction prediction. Our extensive evaluations of various biomedical sources demonstrate that Y-Mol significantly outperforms general-purpose LLMs in discovering lead compounds, predicting molecular properties, and identifying drug interaction events. The source code is available at https://***/r/Y-Mol. © 2024, CC BY.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Exploiting Multi-Decision and Deep Refinement for Ultrasound Image Segmentation

Exploiting Multi-Decision and Deep Refinement for Ultrasound...

引用

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

作者： Wenjing Liu Xuanya Li Kai Hu Xieping Gao Key Laboratory of Intelligent Computing and Information Processing of Ministry of Education Xiangtan University Xiangtan China Baidu Inc Beijing China Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing Hunan Normal University Changsha China

In this paper, we propose a novel convolutional neural network (MDR-Net) for ultrasound image segmentation by exploiting multi-decision and deep refinement of the target. Our MDR-Net consists of two main parts, i.e., a multi-decision module (MDM) and a deep refinement module (DRM). Specifically, the MDM effectively addresses the issue of inconspicuous target regions in ultrasound images by combining multi-scale features and multi-receptive field self-attention to enhance the discriminative representation of features and diagnose feature points multiple times. In addition, to alleviate the problem of blurred boundaries and severe speckle noise, the DRM progressively fuses multi-scale features and makes the fused features interact with higher-level features to refine the target details step by step. Finally, we evaluate the proposed method on two publicly available datasets, namely BUSI and UDIAT. We achieve a Dice of 0.8265 and 0.8827 on the two datasets, which are at least 2% and 1.24% higher than other state-of-the-art ultrasound image segmentation methods.

关键词： Deep learning Image segmentation Visualization Ultrasonic imaging Fuses Neural networks Speckle

来源：评论

学校读者我要写书评

暂无评论

Leveraging Static Relationships for Intra-Type and Inter-Type Message Passing in Video Question Answering

arXiv

引用

arXiv 2025年

作者： Liang, Lili Sun, Guanglu School of Computer Science and Technology Harbin University of Science and Technology Harbin150080 China Heilongjiang Provincial Key Laboratory of Intelligent Information Processing and Application Harbin150080 China

Video Question Answering (VideoQA) is an important research direction in the field of artificial intelligence, enabling machines to understand video content and perform reasoning and answering based on natural language questions. Although methods based on static relationship reasoning have made certain progress, there are still deficiencies in the accuracy of static relationship recognition and representation, and they have not fully utilized the static relationship information in videos for in-depth reasoning and analysis. Therefore, this paper proposes a reasoning method for intra-type and inter-type message passing based on static relationships. This method constructs a dual graph for intra-type message passing reasoning and builds a heterogeneous graph based on static relationships for inter-type message passing reasoning. The intra-type message passing reasoning model captures the neighborhood information of targets and relationships related to the question in the dual graph, updating the dual graph to obtain intra-type clues for answering the question. The inter-type message passing reasoning model captures the neighborhood information of targets and relationships from different categories related to the question in the heterogeneous graph, updating the heterogeneous graph to obtain inter-type clues for answering the question. Finally, the answers are inferred by combining the intra-type and inter-type clues based on static relationships. Experimental results on the ANetQA and Next-QA datasets demonstrate the effectiveness of this method. © 2025, CC BY.

关键词： Question answering

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：