检索结果-内蒙古大学图书馆

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Lucas Ribeiro Hélder P. Oliveira Xiao Hu Tania Pereira INESC TEC INESC TEC - Institute for Systems and Computer Engineering Technology and Science FEUP - Faculty of Engineering University of Porto Porto Portugal INESC TEC - Institute for Systems and Computer Engineering Technology and Science FCUP - Faculty of Science University of Porto Porto Portugal Department of Biomedical Informatics School of Medicine Department of Computer Science College of Arts and Sciences Nell Hodgson Woodruff School of Nursing Emory University Atlanta USA INESC TEC - Institute for Systems and Computer Engineering Technology and Science Porto Portugal

PPG signal is a valuable resource for continuous heart rate monitoring; however, this signal suffers from artifact movements, which is particularly relevant during physical exercise and makes this biomedical signal difficult to use for heart rate detection during those activities. The purpose of this study was to develop learning models to determine heart rate using data from wearables (PPG and acceleration signals) and dealing with noise during physical exercise. Learning models based on CNNs and LSTMs were developed to predict the heart rate. The PPG signal was combined with data from accelerometers trying to overcome the noise movement on the PPG signal. Two datasets were used on this work: the 2015 IEEE Signal Processing Cup (SPC) dataset was used for training and testing, and another dataset was used for validation of the learning model (PPG-DaLiA dataset). The predictions obtained by the learning model represented a mean average error of 7.033±5.376 bpm for the SCP dataset, while a mean average error of 9.520±8.443 bpm for the validation set. The use of acceleration data increases the performance of the learning models on the prediction of the heart rate, showing the benefits of using this source of data to overcome the noise movement problem on the PPG signal. The combination of PPG signal with acceleration data could allow the learning models to use more information regarding the motion artifacts that affect the PPG and improve performance on the physiological event detections, which will largely spread the use of wearables on the healthcare applications for continuous monitor the physiological state allowing early and accurate detection of pathological events.

关键词：

来源：评论

学校读者我要写书评

暂无评论

AdvDisplay: Adversarial Display Assembled by Thermoelectric Cooler for Fooling Thermal Infrared Detectors 39

AdvDisplay: Adversarial Display Assembled by Thermoelectric ...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Li, Hao Wan, Fanggao Su, Yue Wu, Yue Zhang, Mingyang Gong, Maoguo School of Electronic Engineering Xidian University Xi'an710071 China School of Artificial Intelligence Xidian University Xi'an710071 China School of Computer Science and Technology Xidian University Xi'an710071 China Key Laboratory of Collaborative Intelligence Systems Ministry of Education Xi'an710071 China

ISBN: (纸本)157735897X

When the current physical adversarial patches cannot deceive thermal infrared detectors, the existing techniques implement adversarial attacks from scratch, such as digital patch generation, material production, and physical deployment. Besides, it is difficult to finely regulate infrared radiation. To address these issues, this paper designs an adversarial thermal display (AdvDisplay ) by assembling thermoelectric coolers (TECs) as an array. Specifically, to reduce the gap between patches in the physical and digital worlds and decrease the power of AdvDisplay device, heat transfer loss and electric power loss are designed to guide the patch optimization. In addition, a precise temperature control scheme for AdvDisplay is proposed based on proportional-integral-derivative (PID) control. Due to the accurate temperature regulation and the reusability of AdvDisplay, our method is able to improve the attack success rate and the efficiency of physical deployments. Extensive experimental results indicate that the proposed method possesses superior adversarial effectiveness compared to other methods and demonstrates strong robustness in physical attacks. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Thermoelectric refrigeration

来源：评论

学校读者我要写书评

暂无评论

An Effective Online Collaborative Training in Developing Listening Comprehension Skills

引用

computer systems science & engineering 2021年第8期38卷 131-140页

作者： Shakeel Ahmed Munazza Ambreen Muneer Ahmad Abdulellah A.Alaboudi Roobaea Alroobaea NZ Jhanjhi Department of Secondary Instructors Education Faculty of EducationAllama Iqbal Open UniversityIslamabadPakistan Department of Information Systems Faculty of Computer Science&ITUniversiti Malaya50603 Kuala LumpurMalaysia College of Computer Science Shaqra UniversitySaudi Arabia Department of Computer Science College of Computers and Information TechnologyTaif UniversityP.O.Box 11099Taif21944Saudi Arabia School of Computer Science and Engineering SCETaylor’s UniversitySubang Jaya47500Malaysia

The COVID-19 outbreak severely affected formal face-to-face classroom teaching and ***-based online education and training can be a useful measure during the *** the Pakistani educational context,the use of ICT-based online training is generally sporadic and often unavailable,especially for developing English-language instructors’listening comprehension *** major factors affecting availability include insufficient IT resources and infrastructure,a lack of proper online training for speech and listening,instructors with inadequate academic backgrounds,and an unfavorable environment for ICT-based training for listening *** study evaluated the effectiveness of ICT-based training for developing secondary-level English-language instructors’listening comprehension *** this end,collaborative online training was undertaken using random ***,60 private-school instructors in Chakwal District,Pakistan,were randomly selected to receive online-listening training sessions using English *** experimental group achieved significant scores in the posttest ***,there were substantial improvements in the participants’listening skills via online *** the unavailability of face-to-face learning during COVID-19,this study recommends using ICT-based online training to enhance listening comprehension *** policymakers should revise curricula based on online teaching methods and modules.

关键词： COVID-19 online training remote teaching computers in education listening comprehension English language

来源：评论

学校读者我要写书评

暂无评论

Volumetric B1+ field homogenization in 7 Tesla brain MRI using metasurface scattering

arXiv

引用

arXiv 2024年

作者： Yoon, Gyoungsub Yu, Sunkyu Lee, Jongho Park, Namkyoo Photonic Systems Laboratory Department of Electrical and Computer Engineering Seoul National University Seoul08826 Korea Republic of Intelligent Wave Systems Laboratory Department of Electrical and Computer Engineering Seoul National University Seoul08826 Korea Republic of Laboratory for Imaging Science and Technology Department of Electrical and Computer Engineering Seoul National University Seoul Korea Republic of

Ultrahigh field magnetic resonance imaging (UHF MRI) has become an indispensable tool for human brain imaging, offering excellent diagnostic accuracy while avoiding the risks associated with invasive modalities. When the radiofrequency magnetic field of the UHF MRI encounters the multifaceted complexity of the brain—characterized by wavelength-scale, dissipative, and random heterogeneous materials—detrimental mesoscopic challenges such as B1+ field inhomogeneity and local heating arise. Here we develop the metasurface design inspired by scattering theory to achieve the volumetric field homogeneity in the UHF MRI. The method focuses on finding the scattering ansatz systematically and incorporates a pruning technique to achieve the minimum number of participating modes, which guarantees stable practical implementation. Using full-wave analysis of realistic human brain models under a 7 Tesla MRI, we demonstrate more than a twofold improvement in field homogeneity and suppressed local heating, achieving better performance than even the commercial 3 Tesla MRI. The result shows a noninvasive generalization of constant intensity waves in optics, offering a universal methodology applicable to higher Tesla MRI. © 2024, CC BY.

关键词： Magnetic resonance imaging

来源：评论

学校读者我要写书评

暂无评论

Efficient Multi-user Offloading of Personalized Diffusion Models: A DRL-Convex Hybrid Solution

引用

IEEE Transactions on Mobile Computing 2025年

作者： Yang, Wanting Xiong, Zehui Guo, Song Mao, Shiwen Kim, Dong In Debbah, Merouane Singapore University of Technology and Design Pillar of Engineering Systems and Design Singapore Singapore University of Technology and Design Pillar of Information Systems Technology and Design Singapore The Hong Kong University of Science and Technology Department of Computer Science and Engineering Hong Kong Hong Kong Auburn University Department of Electrical and Computer Engineering Auburn36830 United States Sungkyunkwan University Department of Electrical and Computer Engineering Suwon16419 Korea Republic of Khalifa University KU 6G Research Center Department of Computer and Information Engineering Abu Dhabi127788 United Arab Emirates University Paris-Saclay CentraleSupelec Gif-sur-Yvette91192 France

Generative diffusion models like Stable Diffusion are at the forefront of the thriving field of generative models today, celebrated for their robust training methodologies and high-quality photorealistic generation capabilities. These models excel in producing rich content, establishing them as essential tools in the industry. Building on this foundation, the field has seen the rise of personalized content synthesis as a particularly exciting application. However, the large model sizes and iterative nature of inference make it difficult to deploy personalized diffusion models broadly on local devices with heterogeneous computational power. To address this, we propose a novel framework for efficient multi-user offloading of personalized diffusion models. This framework accommodates a variable number of users, each with different computational capabilities, and adapts to the fluctuating computational resources available on edge servers. To enhance computational efficiency and alleviate the storage burden on edge servers, we propose a tailored multi-user hybrid inference approach. This method splits the inference process for each user into two phases, with an optimizable split point. Initially, a cluster-wide model processes low-level semantic information for each user's prompt using batching techniques. Subsequently, users employ their personalized models to refine these details during the later phase of inference. Given the constraints on edge server computational resources and users' preferences for low latency and high accuracy, we model the joint optimization of each user's offloading request handling and split point as an extension of the Generalized Quadratic Assignment Problem (GQAP). Our objective is to maximize a comprehensive metric that balances both latency and accuracy across all users. To solve this NP-hard problem, we transform the GQAP into an adaptive decision sequence, model it as a Markov decision process, and develop a hybrid solution combining dee

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Multimodal Emotion Recognition through Deep Fusion of Audio-Visual Data

Multimodal Emotion Recognition through Deep Fusion of Audio-...

引用

International Conference on computer and Information Technology (ICCIT)

作者： Tamanna Sultana Meskat Jahan Md. Kamal Uddin Yoshinori Kobayashi Mahmudul Hasan Department of Computer Science and Engineering Comilla University Cumilla Bangladesh Department of Computer Science & Telecommunication Engineering NSTU Noakhali Bangladesh Interactive Systems Lab. Saitama University Saitama Japan

The field of emotion recognition in artificial intelligence focuses on enabling machines to comprehend and react to the range of emotions experienced by humans. This paper presents a novel approach that integrates the Convolution Neural Network (CNN) with audio and visual modalities. The study employs the RAVDESS database as a resource to train two distinct models for the analysis of both video and audio data. When it comes to audio pre-processing, advanced signal-processing techniques are applied to extract relevant elements and capture basic acoustic characteristics. A one-dimensional Convolutional Neural Network (CNN) architecture receives the audio data as input, enabling the model to learn complicated patterns and representations from the audio domain. In the context of video pre-processing, sophisticated algorithms are employed to extract essential facial characteristics. In order to capture the changing periods of facial expressions, the video frames are analyzed using a three-dimensional CNN framework following that they have been compressed and converted to grayscale. The fusion technique involves concatenating and extending the outputs of the audio and visual models. The fused features are subsequently sent into a softmax layer, which facilitates the development of a resilient emotion identification system.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Image Caption Generation Related to Object Detection and Colour Recognition Using Transformer-Decoder

Image Caption Generation Related to Object Detection and Col...

引用

International Conference on Computing, Mathematics and engineering Technologies (iCoMET)

作者： Zainab Umair Kamangar Ghulam Mutjaba Shaikh Saif Hassan Nimra Mughal Umair Ayaz Kamangar Computer Science Department Sukkur IBA University Sukkur Pakistan CRAIB Computer Science Department Sukkur IBA University Sukkur Pakistan Computer Systems Engineering Department Sukkur IBA University Sukkur Pakistan

The dependence on digital images is increasing in different fields. i.e, education, business, medicine, or defense, as they are shifting towards the online paradigm. So, there is a dire need for computers and other similar machines to interpret information related to these images and help the users understand the meaning of it. This has been achieved with the help of automatic Image captioning using different prediction models, such as machine learning and deep learning models. However, the problem with the traditional models, especially machine learning models, is that they may not generate a caption that accurately represents that Image. Although deep learning methods are better for generating captions of an image, it is still an open research area that requires a lot of work. Therefore, a model proposed in this research uses transformers with the help of attention layers to encode and decode the image token. Finally, it generates the image caption by identifying the objects along with their colours. The fliker8k and Conceptual Captions datasets are used to train this model, which contains images and captions. The fliker8k contains 8,092 images, each with five captions, and Conceptual Captions contains more than 3 million images, each with one caption. The contribution of this presented work is that it can be utilized by different companies, which require the interpretation of diverse images automatically and the naming of the images to describe some scenario or descriptions related to the images. In the future, the accuracy can be increased by increasing the number of images and captions or incorporating different deep-learning techniques.

关键词： Deep learning Training Image color analysis Text recognition Computational modeling Companies Predictive models

来源：评论

学校读者我要写书评

暂无评论

AdaSin: Enhancing Hard Sample Metrics with Dual Adaptive Penalty for Face Recognition

arXiv

引用

arXiv 2025年

作者： Guo, Qiqi Zheng, Zhuowen Yang, Guanghua Liu, Zhiquan Li, Xiaofan Li, Jianqing Tian, Jinyu Gong, Xueyuan School of Intelligent Systems Science and Engineering Jinan University Guangdong China College of Cyber Security Jinan University Guangdong China School of Computer Science and Engineering Macau University of Science and Technology China

In recent years, the emergence of deep convolutional neural networks has positioned face recognition as a prominent research focus in computer vision. Traditional loss functions, such as margin-based, hard-sample mining-based, and hybrid approaches, have achieved notable performance improvements, with some leveraging curriculum learning to optimize training. However, these methods often fall short in effectively quantifying the difficulty of hard samples. To address this, we propose Adaptive Sine (AdaSin) loss function, which introduces the sine of the angle between a sample’s embedding feature and its ground-truth class center as a novel difficulty metric. This metric enables precise and effective penalization of hard samples. By incorporating curriculum learning, the model dynamically adjusts classification boundaries across different training stages. Unlike previous adaptive-margin loss functions, AdaSin introduce a dual adaptive penalty, applied to both the positive and negative cosine similarities of hard samples. This design imposes stronger constraints, enhancing intra-class compactness and inter-class separability. The combination of the dual adaptive penalty and curriculum learning is guided by a well-designed difficulty metric. It enables the model to focus more effectively on hard samples in later training stages, and lead to the extraction of highly discriminative face features. Extensive experiments across eight benchmarks demonstrate that AdaSin achieves superior accuracy compared to other state-of-the-art methods. Copyright © 2025, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

CASCADING DEEP LEARNING APPROACH FOR IDENTIFYING FACIAL EXPRESSION YOLO METHOD 1

CASCADING DEEP LEARNING APPROACH FOR IDENTIFYING FACIAL EXPR...

引用

1st International Conference on Technologies for Smart Green Connected Society 2021, ICTSGS 2021

作者： Dhanalakshmi, P. Namratha, P. Ramu, Gandikota Prasad, M. Ravi Computer Science and Systems Engineering SVEC Tirupati India Gates Institute of Technology Gooty India Department of CSE Institute of Aeronautical Engineering Hyderabad India S V college of engineering Kadapa India

ISBN: (纸本)9781607685395

Face detection is one of the biggest tasks to find things. Identification is usually the first stage of facial recognition. and identity verification. In recent years in-depth learning algorithms have changed dramatically in object acquisition. These algorithms can usually be divided into two groups, namely two-phase machines like Faster R-CNN or single-phase machines like YOLO. While YOLO and its variants are less accurate than the two-phase detection systems, they outperform other components with wider genes. When faced with standard-sized objects, YOLO works well, but can't get smaller objects. A face recognition system that uses AI (Artificial Intelligence) separates or verifies a person's identity by analyzing their faces. In this project a single neural network predicts binding boxes and class opportunities directly from the full images in a single test. © The Electrochemical Society

关键词： Cascade Deep Learning Face Detection Facial Identification Facial Recognition Yolo Method

来源：评论

学校读者我要写书评

暂无评论

Control when confidence is costly

arXiv

引用

arXiv 2024年

作者： Castillo, Itzel Olivos Schrater, Paul Pitkow, Xaq Department of Computer Science Rice University HoustonTX77005 United States Departments of Computer Science and Psychology University of Minnesota MinneapolisMN55455 United States Departments of Electrical and Computer Engineering and Computer Science Rice University HoustonTX77005 United States Neuroscience Institute Department of Machine Learning Carnegie Mellon University PittsburghPA15213 United States Department of Neuroscience Baylor College of Medicine HoustonTX77030 United States

We develop a version of stochastic control that accounts for computational costs of inference. Past studies identified efficient coding without control, or efficient control that neglects the cost of synthesizing information. Here we combine these concepts into a framework where agents rationally approximate inference for efficient control. Specifically, we study Linear Quadratic Gaussian (LQG) control with an added internal cost on the relative precision of the posterior probability over the world state. This creates a trade-off: an agent can obtain more utility overall by sacrificing some task performance, if doing so saves enough bits during inference. We discover that the rational strategy that solves the joint inference and control problem goes through phase transitions depending on the task demands, switching from a costly but optimal inference to a family of suboptimal inferences related by rotation transformations, each misestimate the stability of the world. In all cases, the agent moves more to think less. This work provides a foundation for a new type of rational computations that could be used by both brains and machines for efficient but computationally constrained control. © 2024, CC BY.

关键词： Stochastic systems

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：