检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Li, Yixiao Yang, Xiaoyuan Fu, Jun Yue, Guanghui Zhou, Wei School of Mathematical Sciences Beihang University China School of Information Sciences University of Science and Technology of China China School of Biomedical Engineering Shenzhen University China School of Computer Science and Informatics Cardiff University United Kingdom

There has emerged a growing interest in exploring efficient quality assessment algorithms for image super-resolution (SR). However, employing deep learning techniques, especially dual-branch algorithms, to automatically evaluate the visual quality of SR images remains challenging. Existing SR image quality assessment (IQA) metrics based on two-stream networks lack interactions between branches. To address this, we propose a novel full-reference IQA (FR-IQA) method for SR images. Specifically, producing SR images and evaluating how close the SR images are to the corresponding HR references are separate processes. Based on this consideration, we construct a deep Bi-directional Attention Network (BiAtten-Net) that dynamically deepens visual attention to distortions in both processes, which aligns well with the human visual system (HVS). Experiments on public SR quality databases demonstrate the superiority of our proposed BiAtten-Net over state-of-the-art quality assessment methods. In addition, the visualization results and ablation study show the effectiveness of bi-directional attention. Copyright © 2024, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Hierarchical Multimodal Decoupling-Fusion Framework for offline Multiple Appropriate Facial Reaction Generation

Hierarchical Multimodal Decoupling-Fusion Framework for offl...

引用

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

作者： Qincheng Lv Xiaofeng Liu Jie Li Rongrong Ni Pujun Xue Siyang Song College of Information Science and Engineering Hohai University Changzhou China College of Artificial Intelligence and Automation Hohai University Changzhou China School of Biomedical Engineering and Informatics Nanjing Medical University Nanjing China School of Microelectronics and Control Engineering Changzhou University Changzhou China School of Computer Science University of Exeter Exeter U.K.

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Facial reactions convey crucial emotional information and coordinating interpersonal relationships in human dyadic interactions. While existing Multiple Appropriate Facial Reaction Generation (MAFRG) methods focus on generating multiple reasonable facial reactions, none of these approaches combines 2D and 3D facial behaviour information nor account for the influence of individuals’ facial identities, leading to inconsistencies in the generated facial reactions and limited capability in capturing subtle variations in facial depth and expression dynamics. This paper proposes a novel Hierarchical Multimodal Decoupling-Fusion (HMDF) framework that decouples 3D facial identity from expression behaviors, eliminating identity-based interference in the reaction generation process, which are integrated with audio-visual features through a cross-attention mechanism. Experiments show that our framework achieved the enhanced diversity and synchrony in the generated facial reactions.

关键词： Three-dimensional displays Interference Signal processing Acoustics Speech processing

来源：评论

学校读者我要写书评

暂无评论

ECRC: Emotion-Causality Recognition in Korean Conversation for GCN

arXiv

引用

arXiv 2024年

作者： Lee, Jun Koo Kim, Moon-Hyun Chung, Tai-Myoung Department of Artificial Intelligence College of Science Sungkyunkwan University Korea Republic of Department of Computing and Informatics College of Science Sungkyunkwan University Korea Republic of Department of Computer Science and Engineering College of Science Sungkyunkwan University Korea Republic of

In this multi-task learning study on simultaneous analysis of emotions and their underlying causes in conversational contexts, deep neural network methods were employed to effectively process and train large labeled datasets. However, these approaches are typically limited to conducting context analyses across the entire corpus because they rely on one of the two methods: word- or sentence-level embedding. The former struggles with polysemy and homonyms, whereas the latter causes information loss when processing long sentences. In this study, we overcome the limitations of previous embeddings by utilizing both word- and sentence-level embeddings. Furthermore, we propose the emotion-causality recognition in conversation (ECRC) model, which is based on a novel graph structure, thereby leveraging the strengths of both embedding methods. This model uniquely integrates the bidirectional long short-term memory (Bi-LSTM) and graph neural network (GCN) models for Korean conversation analysis. Compared with models that rely solely on one embedding method, the proposed model effectively structures abstract concepts, such as language features and relationships, thereby minimizing information loss. To assess model performance, we compared the multi-task learning results of three deep neural network models with varying graph structures. Additionally, we evaluated the proposed model using Korean and English datasets. The experimental results show that the proposed model performs better in emotion and causality multi-task learning (74.62% and 75.30%, respectively) when node and edge characteristics are incorporated into the graph structure. Similar results were recorded for the Korean ECC and Wellness datasets (74.62% and 73.44%, respectively) with 71.35% on the IEMOCAP English dataset. Copyright © 2024, The Authors. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

ADX: IOT-Based Integrated Smart Control System for Prevention of Airborne COVID-19 Spread

ADX: IOT-Based Integrated Smart Control System for Preventio...

引用

International Conference on Information Technology, Information Systems and Electrical engineering (ICITISEE)

作者： Rizqi Sukma Kharisma Kusrini Uyock Anggoro Saputro Mulia Sulistiyono Majid Rahardi Bernadhed Elik Hari Muktafin Faculty of Computer Science Universitas Amikom Yogyakarta Yogvakarta Indonesia Magister of Informatics Engineering Universitas Amikom Yogyakarta Yogyakarta Indonesia

Enclosed spaces are places with a higher potential for spreading the COVID-19 virus. This is because the COVID-19 virus can be carried in the air. Closed space makes the air last longer in the space. Moreover, closed spaces are widely used, such as homes, schools, malls, offices, places of worship, etc. So for closed spaces, serious attention must be given to avoid the spread of the virus. ADX is an IoT-based tool equipped with UVC rays that can kill viruses, including the COVID-19 virus. This tool can be controlled remotely, manually, or on a timer. So that the device can be activated first before the closed space is used. This study also used the ESP8266 microcontroller in supporting the development of IoT-based ADX tools. The results of this study with an accuracy of the IoT ADX system of 96.10% and the average response time is 1.47 seconds.

关键词： COVID-19 Electrical engineering Microcontrollers Aerospace electronics Control systems Time factors Information technology

来源：评论

学校读者我要写书评

暂无评论

SLAH: A Siamese Network Vision Tracking Algorithm Using LSTM Adaptive Head

SLAH: A Siamese Network Vision Tracking Algorithm Using LSTM...

引用

International Conference on Information Science and Technology (ICIST)

作者： Zhitao Liu Xiaonan Luo Dong Wang Fang Li School of Computer Science and Information Security Guilin University of Electronic Technology Guilin China College of Mathematics and Informatics College of Software Engineering South China Agricultural Univerrsity Guangzhou China

In the field of single object tracking, conventional methods often rely on correlation filters or visual image processing. However, these approaches typically focus solely on extracting target object features and consider only the position information in the current frame. They lack integration of contextual and spatial-temporal information, which can limit tracking *** method takes inspiration from single object detection and tracking techniques. It combines efficient model-based object detection with the spatial position of the final target in the image. This approach offers significant advantages in reducing tracking drift (deviation from the target object) and improving tracking accuracy (predicting bounding boxes that closely follow the target).To extract spatial information, we utilize a three-layer ResNet and Convolutional Block Attention Module (CBAM) in conjunction with a siamese network. This combination effectively captures the spatial characteristics of the target object. Additionally, we employ an adaptive processing head with an internal structure of Long Short Term Memory (LSTM) to capture temporal information. By integrating spatial and temporal cues, our method achieves more robust and accurate *** comprehensive comparison and ablation experiments across multiple datasets, we have demonstrated notable improvements with our method. It tightly connects the predicted bounding box with the tracking target and effectively combines spatial and temporal information for precise object tracking.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Learning to Predict Navigational Patterns from Partial Observations

arXiv

引用

arXiv 2023年

作者： Karlsson, Robin Carballo, Alexander Lepe-Salazar, Francisco Fujii, Keisuke Ohtani, Kento Takeda, Kazuya Graduate School of Informatics Nagoya University Aichi Japan Department of Electrical Electronic and Computer Engineering Gifu University Gifu Japan Ludolab Colima Mexico TIER IV Tokyo Japan

Human beings cooperatively navigate rule-constrained environments by adhering to mutually known navigational patterns, which may be represented as directional pathways or road lanes. Inferring these navigational patterns from incompletely observed environments is required for intelligent mobile robots operating in unmapped locations. However, algorithmically defining these navigational patterns is nontrivial. This paper presents the first self-supervised learning (SSL) method for learning to infer navigational patterns in real-world environments from partial observations only. We explain how geometric data augmentation, predictive world modeling, and an information-theoretic regularizer enable our model to predict an unbiased local directional soft lane probability (DSLP) field in the limit of infinite data. We demonstrate how to infer global navigational patterns by fitting a maximum likelihood graph to the DSLP field. Experiments show that our SSL model outperforms two SOTA supervised lane graph prediction models on the nuScenes dataset. We propose our SSL method as a scalable and interpretable continual learning paradigm for navigation by perception. Code is available at https://***/robin-karlsson0/dslp. Copyright © 2023, The Authors. All rights reserved.

关键词： Navigation

来源：评论

学校读者我要写书评

暂无评论

Evaluation of Valued Tolerance Rough Set and Decision Rules Method for WiFi-Based Indoor Localization in Different Environments

Evaluation of Valued Tolerance Rough Set and Decision Rules ...

引用

Proceedings of the International Conference on Intelligent Systems and Network, ICISN 2023

作者： Duong-Bao, Ninh He, Jing Thi, Luong Nguyen Lee, Seon-Woo Nguyen-Huu, Khanh College of Computer Science and Electronic Engineering Hunan University Changsha410082 China Faculty of Mathematics and Informatics Dalat University Dalat66100 Viet Nam Faculty of Information Technology Dalat University Dalat66100 Viet Nam Division of Software Hallym University Chuncheon24252 Korea Republic of Department of Electronics and Telecommunications Dalat University Dalat66100 Viet Nam

ISBN: (纸本)9789819947249

Among various technologies being applied for indoor localization, WiFi has become a common source of information to determine the pedestrian’s position due to the widespread of WiFi access points in indoor environments. However, the fluctuation of the WiFi signals makes it difficult to achieve a good localization result. In this paper, to handle this problem, the valued tolerance rough set and decision rules method (VTRS-DR), which is firstly registered to WiFi fingerprinting-based localization, will be implemented and evaluated in big and complicated environments using two public datasets. The first one was conducted by a subject using a smartphone at a multi-floor library for several months. Furthermore, to evaluate the localization accuracy when WiFi data was collected from different pedestrians as well as different smartphones, a crowdsourced WiFi fingerprinting dataset was utilized. From the deep analyses of localization results, the VTRS-DR method shows high accuracy and high robustness when testing in different environments with a mean error of 3.84 m, which is 27.87% lower than other compared methods. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Smartphones

来源：评论

学校读者我要写书评

暂无评论

GPU Parallel Computing for Detection and Classification Object in Robot Soccer ERSOW

GPU Parallel Computing for Detection and Classification Obje...

引用

2020 International Electronics Symposium, IES 2020

作者： Anwar, Khoirul Haq, Muhammad Abdul Kurnianto Wibowo, Iwan Bachtiar, Mochamad Mobed Politeknik Elektronika Negeri Surabaya Department of Informatics and Computer Engineering Surabaya Indonesia

ISBN: (纸本)9781728195308

ERSOW robot soccer that participated in the Indonesian Wheeled Robot Soccer Contest, has many abilities such as object detection and classification, control and navigation system, self-localization and mapping, and also real-Time communication between each other. This research focusing on object detection and classification on the robots as one of important processes to provide main data sources for all further actions. Therefore, this process takes longer computation time to detect and classify multiple objects, to improve detection and classification performance speed without sacrifice the accuracy value, we proposed GPU parallel computing on offline training phase and online inference phase. In the offline training phase, the neural network model can be trained in parallel processes using selected GPU hardware. As a result of training, we can transfer learning the model knowledge to another host. The experiments on the NVIDIA Jetson AGX Xavier board show that the custom model as the result of the offline training phase achieves more than 30 fps and pre-Trained model SSD-MobileNet-v2 achieve more than 99 fps. © 2020 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Detecting multi-pose masked face using adaptive boosting and cascade classifier

Detecting multi-pose masked face using adaptive boosting and...

引用

2020 International Electronics Symposium, IES 2020

作者： Bayu Dewantara, Bima Sena Twinda Rhamadhaningrum, Dhiska Politeknik Elektronika Negeri Surabaya Dept of Informatics and Computer Engineering Surabaya Indonesia

ISBN: (纸本)9781728195308

This paper presents the use of the Adaptive Boosting and Cascade Classifier based method to detect someone wearing a mask or not on various facial poses. In general, masks are used to protect the nose and mouth to prevent dirt or bacteria or virus from entering the respiratory tract Simply stated, a person is said to use a mask well if the nose and mouth areas are not visually visible on the face. OpenCV has already provided a model for human face detection for both frontal and profile faces. These models are good for detecting human frontal faces and profile faces, but not to detect people with masks. To deals with multi-pose people's faces using masks, we have to train new nose and mouth-based Adaptive Boosting and Cascade Classifies based model on various face poses independently to achieve our expectation. The Caspeal face database and AISL face database are used to train the Haar-like feature, LBP feature, and HOG feature based Adaptive Boostings and Cascade Classifiers. Based on the results of experiments, Haar-like feature outperforms LBP feature and HOG feature by achieving the best detection accuracy of 86.9%. On the other hand, LBP features outperforms Haar-like feature and HOG feature in computation time by achieving less than 30 msec from image loading until showing the detection result. © 2020 IEEE.

关键词： Adaptive boosting

来源：评论

学校读者我要写书评

暂无评论

Investigating the Predictive Reproducibility of Federated Graph Neural Networks using Medical Datasets

arXiv

引用

arXiv 2022年

作者： Balık, Mehmet Yigit Rekik, Arwa Rekik, Islem BASIRA Lab Faculty of Computer and Informatics Engineering Istanbul Technical University Istanbul Turkey Faculty of Medicine of Sousse Sousse Tunisia

Graph neural networks (GNNs) have achieved extraordinary enhancements in various areas including the fields medical imaging and network neuroscience where they displayed a high accuracy in diagnosing challenging neurological disorders such as autism. In the face of medical data scarcity and high-privacy, training such data-hungry models remains challenging. Federated learning brings an efficient solution to this issue by allowing to train models on multiple datasets, collected independently by different hospitals, in fully data-preserving manner. Although both state-of-the-art GNNs and federated learning techniques focus on boosting classification accuracy, they overlook a critical unsolved problem: investigating the reproducibility of the most discriminative biomarkers (i.e., features) selected by the GNN models within a federated learning paradigm. Quantifying the reproducibility of a predictive medical model against perturbations of training and testing data distributions presents one of the biggest hurdles to overcome in developing translational clinical applications. To the best of our knowledge, this presents the first work investigating the reproducibility of federated GNN models with application to classifying medical imaging and brain connectivity datasets. We evaluated our framework using various GNN models trained on medical imaging and connectomic datasets. More importantly, we showed that federated learning boosts both the accuracy and reproducibility of GNN models in such medical learning tasks. Our source code is available at https://***/basiralab/reproducibleFedGNN. © 2022, CC BY-NC-ND.

关键词： Graph neural networks

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：