检索结果-内蒙古大学图书馆

IEEE International Conference on Multimedia and Expo (ICME)

作者： Yi Li Xin Xie Haiyan Fu Xiangyang Luo Yanqing Guo School of Artificial Intelligence Dalian University of Technology Dalian China School of Information and Communication Engineering Dalian University of Technology Dalian China State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China

Due to the limitation of spatial receptive field, it is challenging for CNN-based style transfer methods to capture rich and long-range semantic concepts in artworks. Though the transformer provides a fresh solution by considering long-range dependencies, it suffers from the heavy burdens of parameter scale and computation cost especially in vision tasks. In this paper, we design a compact transformer architecture AdaFormer to address the problem. The model scale shrinks about 20% compared to state-of-the-art transformer for style transfer. Furthermore, we explore the adaptive style transfer by letting the content to select the detailed style element automatically and adaptively, which encourages the output to be both appealing and reasonable. We evaluate AdaFormer comprehensively in the experiments and the results have shown the effectiveness and superiority of our approach compared to existing artistic methods. Diverse plausible stylized images are obtained with better content preservation, more convincing style sfumato and lower computation complexity.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Focus, Distinguish, and Prompt: Unleashing CLIP for Efficient and Flexible Scene Text Retrieval

arXiv

引用

arXiv 2024年

作者： Zeng, Gangyan Zhang, Yuan Wei, Jin Yang, Dongbao Zhang, Peng Gao, Yiwen Qin, Xugong Zhou, Yu School of Cyber Science and Engineering Nanjing University of Science and Technology China State Key Laboratory of Media Convergence and Communication Communication University of China China Institute of Information Engineering Chinese Academy of Sciences China Laboratory for Advanced Computing and Intelligence Engineering China TMCC College of Computer Science Nankai University China

Scene text retrieval aims to find all images containing the query text from an image gallery. Current efforts tend to adopt an Optical Character Recognition (OCR) pipeline, which requires complicated text detection and/or recognition processes, resulting in inefficient and inflexible retrieval. Different from them, in this work we propose to explore the intrinsic potential of Contrastive Language-Image Pre-training (CLIP) for OCR-free scene text retrieval. Through empirical analysis, we observe that the main challenges of CLIP as a text retriever are: 1) limited text perceptual scale, and 2) entangled visual-semantic concepts. To this end, a novel model termed FDP (Focus, Distinguish, and Prompt) is developed. FDP first focuses on scene text via shifting the attention to the text area and probing the hidden text knowledge, and then divides the query text into content word and function word for processing, in which a semantic-aware prompting scheme and a distracted queries assistance module are utilized. Extensive experiments show that FDP significantly enhances the inference speed while achieving better or competitive retrieval accuracy compared to existing methods. Notably, on the IIIT-STR benchmark, FDP surpasses the state-of-the-art model by 4.37% with a 4 times faster speed. Furthermore, additional experiments under phrase-level and attribute-aware scene text retrieval settings validate FDP’s particular advantages in handling diverse forms of query text. The source code will be publicly available at https://***/Gyann-z/FDP. © 2024, CC BY-NC-SA.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

通过表面终止化学设计实现高度可逆的锌金属阳极

引用

Science Bulletin 2023年第23期68卷 2993-3002,M0005页

作者：刘宇陈树麟袁豪熊方宇刘琴安永康张建勇吴璐孙建国 Yong-Wei Zhang 安琴友 John Wang State Key Laboratory of Advanced Technology for Materials Synthesis and Processing Wuhan University of TechnologyWuhan 430070China Department of Materials Science and Engineering National University of SingaporeSingapore 117574Singapore Institute of High Performance Computing Agency for ScienceTechnology and Research(A*Star)Singapore 138632Singapore Institute of Materials Research and Engineering(IMRE) Agency for ScienceTechnology and Research(A*Star)Singapore 138634Singapore

锌表面自然形成的非均匀氧化层(ZnO)会与电解液发生反应形成Zn_(4)(OH)_(6)SO_(4)·xH_(2)O(ZSH)副产物,导致锌负极表面非均匀沉积,从而阻碍水性锌离子电池实际应用.为了排除氧化层分布形式的影响,本文利用表面终止化学设计策略在... 详细信息

锌表面自然形成的非均匀氧化层(ZnO)会与电解液发生反应形成Zn_(4)(OH)_(6)SO_(4)·xH_(2)O(ZSH)副产物,导致锌负极表面非均匀沉积,从而阻碍水性锌离子电池实际应用.为了排除氧化层分布形式的影响,本文利用表面终止化学设计策略在锌负极表面制备了一层完全氧化的ZnO层,研究发现锌负极均匀分布的ZnO层会原位演化形成致密的ZSH功能层.均匀且致密的ZSH能够有效隔绝电解液的腐蚀,从而有效提高锌负极界面的稳定性.此外,ZSH(002)晶面边缘的不饱和悬空键(如-O)更容易和Zn(H_(2)O)_(6)^(2+)结构中的锌离子形成强Zn-O键,可以促进Zn(H_(2)O)_(6)^(2+)的脱溶剂化和锌离子沿边扩散,实现了高度可逆的锌沉积行为并限制了副反应.基于此,本文提出与“表面脱水-晶体通道转移”机制明显不同的“边缘脱水-沿边缘转移”锌离子输运新机制.更重要的是,基于此组装的全电池显示出高循环稳定性(超过1000次循环)和高库仑效率(99.07%),验证了其潜在的实际应用价值.

关键词： Zn-O termination Zn Anode Solid electrolyte interphase Dendrite-free Zn ion battery

来源：评论

学校读者我要写书评

暂无评论

Self-Attention Prediction Correction with Channel Suppression for Weakly-Supervised Semantic Segmentation

Self-Attention Prediction Correction with Channel Suppressio...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Guoying Sun Meng Yang School of Computer Science and Engineering Sun Yat-sen University Guangzhou China State Key Laboratory of Integrated Services Networks Xidian University Key Laboratory of Machine Intelligence and Advanced Computing Sun Yat-sen University Ministry of Education China

Single-stage weakly-supervised semantic segmentation (WSSS) with image-level labels has become a new research hotspot in the community for its lower cost and higher training efficiency. However, the pseudo label of WSSS generally suffers from somewhat noise, which limits the segmentation performance. In this paper, to explore the integral foreground activation, we propose the Channel Suppression (CS) module for preventing only activating the most discriminative regions, thereby improving the initial pseudo labels. To rectify the in-correct prediction, we explore the Self-Attention Prediction Correction (SAPC) module, which adaptively generates the category-wise prediction rectification weights. After extensive experiments, the proposed efficient single-stage framework achieves excellent performance with 67.6% mIoU and 39.9% mIoU on PASCAL VOC 2012 and MS COCO 2014 datasets, significantly exceeding several recent single-stage methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Security Protocol Analysis Tool SmartVerif

Research on Security Protocol Analysis Tool SmartVerif

引用

2021 International Conference on Computer Technology, Information engineering and Electron Materials, CTIEEM 2021

作者： Lu, Siqi Han, Qingdi Miao, Xuyang Liu, Yubo State Key Laboratory of Mathematical Engineering and Advanced Computing Henan Zhengzhou450000 China Henan Key Laboratory of Network Cryptography Technology Henan Zhengzhou450000 China Department of Information Science and Engineering Ocean University of China Shandong Qingdao266100 China

Security protocols have been designed to protect the security of the network. However, many security protocols cannot guarantee absolute security in real applications. Therefore, security tests of the network protocol become particularly important. In this paper, firstly, we introduce SmartVerif, which is the first formal analysis tool to automatically verify the security of protocols through dynamic strategies. And then, we use SmartVerif to verify the pseudo-randomness of the encapsulated key of the Two-Pass AKE protocol, which was proposed by Liu's in ASIACRYPT in 2020. Finally, we summary our work and show some limitations of SmartVerif. At the same time, we also point out the direction for future improvement of SmartVerif. © Content from this work may be used under the terms of the Creative Commons Attribution 3.0 Licence.

关键词： Network security

来源：评论

学校读者我要写书评

暂无评论

Multiple Images Steganography of JPEG Images Based on Optimal Payload Distribution

引用

Computer Modeling in engineering & Sciences 2020年第10期125卷 417-436页

作者： Yang Pei Xiangyang Luo Yi Zhang Liyan Zhu Zhong Yuan Network Security Research Institute Zhengzhou UniversityZhengzhou450000China State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou450000China

Multiple images steganography refers to hiding secret messages in multiple natural images to minimize the leakage of secret messages during ***,the main multiple images steganography algorithms mainly distribute the payloads as sparsely as possible inmultiple cover images to improve the detection error rate of stego *** order to enable the payloads to be accurately and efficiently distributed in each cover image,this paper proposes a multiple images steganography for JPEG images based on optimal payload ***,the algorithm uses the principle of dynamic programming to redistribute the payloads of the cover images to reduce the time required in the process of payloads ***,by reducing the difference between the features of the cover images and the stego images to increase the detection error rate of the stego ***,this paper uses a data decomposition mechanism based on Vandermonde *** if part of the data is lost during the transmission of the secret messages,as long as the data loss rate is less than the data redundancy rate,the original secret messages can be *** results show that the method proposed in this paper improves the efficiency of payloads distribution compared with existing multiple images *** the same time,the algorithm can achieve the optimal payload distribution of multiple images steganography to improve the anti-statistical detection performance of stego images.

关键词： Multiple images steganography payloads distribution dynamic programming messages recovery

来源：评论

学校读者我要写书评

暂无评论

A Survey on Neural Machine Translation Applied to Sign Language Generation 3

A Survey on Neural Machine Translation Applied to Sign Langu...

引用

3rd International Conference on Applied Machine Learning, ICAML 2021

作者： Zhang, Yue Cao, Lihong State Key Laboratory of Media Convergence and Communication Communication University of China Beijing100024 China Neuroscience and Intelligent Media Institute Communication University of China Beijing100024 China State Key Laboratory of Mathematical Engineering and Advanced Computing Wuxi214125 China

ISBN: (纸本)9781665421256

Sign languages are the primary medium of communication for the deaf and hearing-impaired community, and the number of human beings with hearing impairment is increasing at a rapid rate. There is a signifîcant communication barrier between the hearing-impaired and normal-hearing communities. One solution to this issue is to offload the effort of text imderstanding for the deaf and hearing-impaired community in the form of Sign Language Generation (SLG). In this paper, first of all, we systematically review the mainstream Neural Machine Translation (NMT)based SLG approaches. Specifically, according to their model architectures, we divide them into two categories: the Recurrent Neural Networks (RNNs)-based and Transformer-based models, and discuss their advantages and limitations. Next, we introduce the existing publicly available datasets and the data pre pros ses sing methods in the field of SLG. Then, we summarize the performance evaluation methods for NMT-based SLG models. Finally, we offer two possible future research directions for better development in the field of SLG. © 2021 IEEE

关键词： Neural machine translation

来源：评论

学校读者我要写书评

暂无评论

Shields for Digital Images: A Watermarking Method with KAN Block and Simulation-Enhanced Noise Pool to Resist Screen-Camera Attacks

引用

IEEE Transactions on Circuits and Systems for Video Technology 2025年

作者： Guo, Daidou Qin, Chuan Luo, Xiangyang Feng, Guorui Zhang, Xinpeng University of Shanghai for Science and Technology School of Optical-Electrical and Computer Engineering Shanghai 200093 China State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou 450000 China Henan Key Laboratory of Cyberspace Situation Awareness Zhengzhou 450001 China Shanghai University School of Communication and Information Engineering Shanghai 200444 China

To address the issues of privacy leakage and copyright infringement in screen-camera scenarios, we propose a robust image watermarking method, which incorporates kolmogorov-arnold network (KAN) blocks and a simulation-enhanced noise pool to resist screen-camera noise attacks. Specifically, we first modify the traditional convolutional blocks for processing high-dimensional features in the U-Net-based encoder to KAN blocks. This operation enhances the ability of encoder to model nonlinear relationships between complex features, while preserving global structure of the original image and minimizing the damage to local details caused by watermark embedding, thereby improving visual quality of the watermarked image. Additionally, to enhance the robustness of the proposed method against complex screen-camera noise attacks, a simulation-enhanced noise pool containing mathematical models and a deep noise simulation network, called NSim-Net, is designed. Especially, in NSim-Net, adversarial training between the simulator based on the improved U-Net and the discriminator based on PatchGAN effectively improves the ability to simulate complex noise. Experimental results demonstrate that, compared to typical screen-camera resilient watermarking methods, the watermarked image generated by the proposed method achieves a maximum peak signal-to-noise ratio (PSNR) improvement of 4.78 dB. Furthermore, based on our simulation-enhanced noise pool, the watermark extraction accuracy of the proposed method exceeds 98% under various screen-camera noise attacks. © 1991-2012 IEEE.

关键词： Imperceptibility Robustness Screen-Camera Resilient Watermarking Simulation-Enhanced Noise Pool

来源：评论

学校读者我要写书评

暂无评论

Low dose PET image reconstruction using diffusion sampling with posterior iterative correction

Low dose PET image reconstruction using diffusion sampling w...

引用

IEEE Symposium on Nuclear Science (NSS/MIC)

作者： B. Qi R. Hu D. Wu N. Guo Q. Li H. Liu Department of Optical Engineering the State Key Laboratory of Modern Optical Instrumentation Zhejiang University Hangzhou China Massachusetts General Hospital/Harvard Medical School The Center for Advanced Medical Computing and Analysis Boston Massachusetts United States of America

ISBN: (数字)9798350388152

ISBN: (纸本)9798350388169

Low dose PET image reconstruction is a crucial technique in functional imaging, addressing both patient safety and diagnostic needs by minimizing radioactive exposure. However, as dosage is reduced, the increasing noise makes reconstruction challenging. This study leverages a diffusion model in a pre-training mode to learn PET image priors from existing PET data, establishing a sampling framework for low-dose PET reconstruction. By constraining the sampling process with a data fidelity correction term through iterative reconstruction and implementing posterior iterative correction, we achieve 50x down-sampled PET reconstruction without the need for training with paired data. Experiments demonstrate that our method achieves excellent performance compared with MLEM, EMTV and DeepPET method.

关键词： Training Microwave integrated circuits Semiconductor detectors Noise Diffusion models Safety Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Exploration and research of blended learning in the "principles of computer organization" 2

Exploration and research of blended learning in the "princip...

引用

2nd International Conference on Computer Science and Educational Informatization, CSEI 2020

作者： Hou, Yifan Wang, Wei Liu, Fudong Shan, Zheng Wang, Junchao Wang, Lixin State Key Laboratory of Mathematical Engineering and Advanced Computing Information Engineering University Zhengzhou China

ISBN: (纸本)9781728170077

The'Principles of Computer Organization' is the core basic course for the construction of computer hardware courses. In this paper, it analyzes the characteristics of the 'Principles of Computer Organization 'Aiming at the challenges faced by the traditional teaching. This paper makes a beneficial exploration on the teaching reform of the Principle of Computer Composition based on the blended learning mode from the aspects of teaching content, teaching method transformation, interactive communication, evaluation and assessment, etc. This paper proposes a framework for self-learning ability training based on a blended learning model, which is applied and formed in the course of principles. Teaching practice shows that the reform of teaching mode and teaching practice not only stimulates learning interest and motivation of students, but also improves their self-learning ability and practical ability. © 2020 IEEE.

关键词： Teaching

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：