检索结果-内蒙古大学图书馆

5th International Conference on Advanced Research in Computing, ICARC 2025

作者： Yeshani, R.K.B. Chiththranjan, Charith University of Moratuwa Dept. of Computer Science and Eng. Sri Lanka

ISBN: (纸本)9798331530983

This work presents an innovative approach to image-to-image translation, focusing on converting colour pencil images, grayscale pencil images, and grayscale images into RGB images using the Pix2Pix GAN. Building upon the established Pix2Pix framework, which utilizes a U-Net based architecture with skip connections for the generator and a Patch-GAN discriminator. This aims to refine the translation process, addressing the images' intricate details and colour dynamics. Empirical results demonstrate that this generator architecture significantly outper- forms the U-Net architecture in producing more vivid, detailed and accurate RGB images. By leveraging this method, the study not only contributes to the body of knowledge in neural network applications for artistic image translation but also opens new avenues for further researches in image translations. This work underscores the potential of targeted architectural modifications within GAN frameworks to achieve image translation outcomes, thereby setting a precedent for future explorations in the field. © 2025 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Nearly Lossless Zero Shot Learning for Robot Navigation Using Declarative Re-Planning

Nearly Lossless Zero Shot Learning for Robot Navigation Usin...

引用

2025 IEEE SoutheastCon, SoutheastCon 2025

作者： Hemming, Nathan Menon, Vineetha University of Alabama Huntsville Dept. of Computer Science Huntsville United States

ISBN: (纸本)9798331504847

Navigation for mobile robots is critical for the deployment and use of robots. Without the ability to interact with the world a robot cannot bring to bear its utility. This work further explores the properties of a mobile robot navigation method called declarative re-planning (DRP). DRP is a hybrid method that leverages classical path planning and robot controllers combined with machine learning (ML) based local path planner for navigation. This work demonstrates how DRP is effective at zero shot transfer learning. DRP is able to train a navigation policy on one robot, and be used on other robots while remaining equally effective at reaching its goal. The other robots do not need to have the same properties, actuators, or number of actuators. DRP also enables simulation acceleration by allowing lower fidelity trajectory calculations to be used without effecting the outcome of the navigation trajectory. © 2025 IEEE.

关键词： Mobile robots

来源：评论

学校读者我要写书评

暂无评论

Advanced Decision-Making for Autonomous Vehicles Utilizing Deep Q Networks

Advanced Decision-Making for Autonomous Vehicles Utilizing D...

引用

2025 International Conference on Emerging Technologies for Intelligent Systems, ETIS 2025

作者： Britto, Akash Ij Abhishek, S. Dhanush, S. Salim, Soja Soniya, B. Sct College of Engineering Dept. of Computer Science Engineering Thiruvananthapuram India

ISBN: (纸本)9798331507541

Autonomous vehicles are poised to revolutionize the transportation industry by offering safer and more efficient navigation in dynamic environments. A critical challenge is managing interactions with other vehicles, pedestrians, and road obstacles, which complicates framing autonomous driving as a supervised learning problem. This study leverages Deep Q-Networks (DQN), a reinforcement learning technique, to control steering, throttle, and braking in autonomous vehicles. The DQN approach enables vehicles to autonomously learn and improve driving strategies, thereby enhancing their ability to navigate complex traffic scenarios effectively. Our objective is to train a model using a Deep Reinforcement Learning (DRL) algorithm, guiding the vehicle to adhere to a predetermined itinerary while adapting to real-time traffic conditions. This work emphasizes the potential of reinforcement learning for enabling continuous learning and adaptation in complex traffic scenarios, laying the groundwork for more reliable and adaptive autonomous driving systems. © 2025 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A New Three-phase Multi-mode AC/DC LLC Converter with Output-controlled Active Rectifier (with V2G and G2V functions) For Fast DC Charging Application 14

A New Three-phase Multi-mode AC/DC LLC Converter with Output...

引用

14th Annual IEEE Applied Power Electronics Conference and Exposition, APEC 2025

作者： Xia, Xiaoyi Lam, John York University Dept. of Electrical Engineering and Computer Science Toronto Canada

ISBN: (纸本)9798331516116

This paper proposes a new three-phase multi-mode AC/DC LLC resonant converter with an output-controlled active rectifier for electric vehicle (EV) fast DC charging applications. In the proposed approach, two low-frequency bidirectional switches at the primary side in each phase of the converter allow vehicle-to-grid (V2G, i.e. DC/AC mode) and grid-to-vehicle (G2V, i.e. AC/DC mode) modes. When the proposed converter operates in AC/DC mode, each phase consists of an integrated bridgeless boost power factor corrector and an LLC resonant converter at the primary side. Output voltage regulation is managed by switches on the high frequency transformer's secondary side output rectifier in each phase, with soft-switching operation achieved in all the switches. When the proposed three-phase converter operates in DC/AC mode, each phase consists of a half-bridge DC/DC resonant converter cascaded with a half-bridge grid-side inverter. The operation of the proposed converter is explained in this paper. Results from a 20kW, 480VLLrms/350Vdc, 130kHz design, and a hardware test on a 130kHz, 250V-output proof-of-concept prototype are presented to validate the functionality of the proposed converter. © 2025 IEEE.

关键词： Zero voltage switching

来源：评论

学校读者我要写书评

暂无评论

A Machine Learning Approach for Predicting Personality Types from Text Patterns Using Natural Language Processing(NLP) 8

A Machine Learning Approach for Predicting Personality Types...

引用

8th International Conference on Electronics, Materials Engineering and Nano-Technology, IEMENTech 2025

作者： Ganguly, Kinshuk Saha, Subhajit Mondal, Anupam Dey, Monalisa Institute of Engineering & Management Dept. of Computer Science & Engineering Kolkata India

ISBN: (纸本)9798331520762

This study explores the use of Natural Language Processing (NLP) approach to predict Myers-Briggs Type Indicator (MBTI) personality types from Twitter texts. Initially, we approached the problem as a 16-class classification task, but encountered significant challenges due to the lack of instances representing certain class labels adequately. To address this, we decomposed the MBTI types into four dimensions and trained individual classifiers for each dimension. Leveraging the best performing classifiers, we constructed a stacking ensemble model to predict personality classes corresponding to each dimension. The results demonstrate improved accuracy compared to the initial 16-class classification approach. Additionally, the performance of ensemble model is found to be better than the individual classifiers. Thus, ensemble approach paired with the separate dimension based classification method can be used for effective prediction of personality types for psychological profiling. © 2025 IEEE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Fine-Grained Object Detection with Remote-Sensing data Using Optimized YOLO-based Models

Fine-Grained Object Detection with Remote-Sensing Data Using...

引用

2025 IEEE SoutheastCon, SoutheastCon 2025

作者： Andrianarivony, Henintsoa S. Akhloufi, Moulay A. Université de Moncton Dept. of Computer Science MonctonNB Canada

ISBN: (纸本)9798331504847

Remote sensing data is used in various fields, such as environmental monitoring, urban planning, agriculture, and disaster management. Advances in satellite technology permit the creation of high-quality data, including high-resolution remote-sensing images. However, classical techniques struggle to process and extract information from large quantities and high-resolution remote-sensing data. In this context, with the advancement in computing power, deep learning has become a powerful tool for extracting features and processing remote-sensing data. Despite these advancements, remote-sensing object detection faces challenges related to the diversity of object types, variations in scale and resolution, and the presence of occlusion, which can hinder the accuracy and robustness of detection models. This paper explores YOLO-based architectures for horizontal and oriented bounding box detection in fine-grained object detection using the FAIR1M dataset. FAIR1M provides high-resolution satellite images with 5 object categories and 37 object sub-categories. The best model in horizontal object detection is the YOLOv9e, achieving a mAP50 of 44.6 %. The best model in oriented object detection is a pre-trained model using a custom weighted data loader, achieving a mAP50 of 40.5 %. We further analyze the strengths and limitations of these techniques for fine-grained remote-sensing object detection and highlight the contributions to improving the models' performances depending on the application. We then conclude and give directions for future work. © 2025 IEEE.

关键词： Disaster prevention

来源：评论

学校读者我要写书评

暂无评论

Segmentation of Impervious Areas from High Altitude Aerial Images using a Regularized UNet Framework: ∗ Note: First two authors contributed equally and are sorted by last name

Segmentation of Impervious Areas from High Altitude Aerial I...

引用

2025 IEEE SoutheastCon, SoutheastCon 2025

作者： Athmakur, Srivani Shakri, Shania Mukherjee, Chaity Banerjee University of Alabama in Huntsville Dept. of Computer Science Huntsville United States

ISBN: (纸本)9798331504847

In this paper we introduce a novel dataset of high altitude aerial images with manual annotations for impervious surfaces from a predominantly urban area. The dataset also contains a small amount of hand annotations for man-made pervious surfaces, such as unpaved road and driveways. Each aerial image also has unlabeled regions that are neither impervious nor man-made pervious. A pervious surface is defined as one that allows water to pass through whereas an impervious one does not allow water to percolate. After introducing the dataset, we study the problem of automatically categorizing the regions in an aerial image into their respective categories. To this end we explore three directions: 1) predicting impervious, pervious and unlabeled (background), 2) predicting impervious and the combined class of unlabeled and man-made pervious and 3) predicting impervious and background excluding man-made pervious regions. For each of these problems, we used a custom built UNet architecture with dropout layers. We report the results from the same in the form of classification masks. The results were verified by manually aligning the predicted mask with the ground truth mask for impervious, man-made pervious, and background regions. Our experiments and results establish the efficacy of using deep learning techniques for the problem of automatically identifying impervious surfaces from high-resolution aerial imagery, especially in complex urban areas. © 2025 IEEE.

关键词： Aerial photography

来源：评论

学校读者我要写书评

暂无评论

LeXNet++: Layer-wise eXplainable ResUNet++ framework for segmentation of colorectal polyp cancer images

引用

Neural Computing and Applications 2025年第1期37卷 213-229页

作者： Das, Surajit Khan, Soumya Suvra Sengupta, Diganta Debashis, De Dept. of Information Technology Meghnad Saha Institute of Technology Behind Urbana Complex Near Ruby General Hospital Anandapur Rd Uchhepota Kolkata700150 India Dept. of Computer Science & Engineering Meghnad Saha Institute of Technology Behind Urbana Complex Near Ruby General Hospital Anandapur Rd Uchhepota Kolkata700150 India Dept. of Computer Science & Engineering Heritage Institute of Technology Chowbaga Road West Bengal Anandapur Kolkata700107 India Dept. of Computer Science & Engineering Maulana Abul Kalam Azad University of Technology West Bengal NH-12 Simhat Haringhata West Bengal Nadia741249 India

Colorectal polyps are benign lesions that develop in the colon and can progress to cancer if left untreated. Clinical observations from medical images are often preferred over computational results due to the lack of trust in the machine learning models, thereby posing serious challenge for the explainability of the results. In order to computationally diagnose colorectal polyps from cancerous images and explain the results, we propose a Layer-wise eXplainable ResUNet++ (LeXNet++) framework for segmentation of the cancerous images, followed by layer-wise explanation of the results. We utilize a publicly accessible dataset that contains of 612 raw images with a resolution of 256×256×3 and an additional 612 clinically annotated and labeled images with a resolution of 256×256×1, which includes the infected region. The LeXNet++ framework comprises of three components—encoder, decoder and the bridge. The encoder and the decoder components each comprise of four layers. Each of the four layers in the encoder and the decoder comprises of 14 and 11 internal sub-layers, respectively. Among the sub-layers of the encoder and the decoder, there are three 3×3 convolutional layers with an additional 3×3 convolution-transpose layer in the decoder. The output of each of the sub-layers has been explained through heatmap generation after each iteration which have been further explained. The encoder and the decoder are connected by the bridge which comprises of three sub-layers. The results obtained from these three sub-layers have also been explained to inculcate trust in the findings. In this study, we have used three models to segment the images, namely UNet, ResUNet, and proposed LeXNet++. LeXNet++ exhibited the best result among the three models in terms of performance;hence, only LeXNet++ was explained layer-wise. Apart from explanation of the results fetched in this study, the performance of the proposed explainable model has been observed to be 2% greater than the existing poly

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Recognition of American sign language using modified deep residual CNN with modified canny edge segmentation

引用

Multimedia Tools and Applications 2025年 1-28页

作者： Hariharan, U. Devarajan, Devikanniga Kumar, P. Santhosh Rajkumar, K. Meena, M. Akilan, T. Department of Computer Science and Engineering Apex Institute of Technology Chandigarh University Punjab Mohali140413 India Department of Computer Science and Engineering GITAM University Karnataka Bengaluru561203 India Department of Information Technology Faculty of Engineering and technology SRM Institute of Science and Technology Ramapuram Tamil Nadu Chennai600089 India Department of Data Science and Business Systems School of Computing Faculty of Engineering and Technology SRM Institute of Science and Technology Kattankulathur Tamil Nadu Chennai603203 India Department of Computer Science and Engineering Sri Sairam Engineering College Tamil Nadu Tambaram602109 India Research Coordinator Dept. of IQAC Amity University Kolkata West Bengal 700135 India

American Sign Language (ASL) recognition aims to recognize hand gestures, and it is a crucial solution to communicating between the deaf community and hearing people. However, existing sign language recognition algorithms still have some drawbacks, such as difficulty recognizing hand movements and low recognition accuracy for most sign language recognition. A Modified Convolutional Neural Network (MCNN) deep residual 101 classifier-based American Sign Language identification system has been developed to address this problem. There are three main parts presents in the method. The first part is pre-processing the images to remove the noise, enhance the contrast of the picture, adjust the contrast level and smoothen the picture using various filters. The second part is the segmentation, and it’s used to partition the image using a modified canny edge detection method by removing all weak edges present in the image. Finally, classification will be done using the Modified CNN deep residual 101 classifiers. By classifying the image, the American Sign Language is accurately identified. This process is conducted through images. The outcome shows that the suggested approach has a 0.97% accuracy and a 0.05% False Positive Rate. Other CNN architectures such as resNet 50 and resNet 18 reached 0.90% and 0.80% accuracy, respectively, which is lower than our proposed method. In addition, the resNet 101 classifier effectively recognizes the difficult hand gestures through the image data. It obtains high recognition accuracy for 36 signs is 0 to 9 numbers and a to z alphabets from American Sign Language. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A glimpse ahead: Forecasting 5.5-s human vigilance for enhanced safety in Industry 5.0

引用

Biocybernetics and Biomedical Engineering 2025年第2期45卷 258-268页

作者： Cinquetti, Ettore Siviero, Ilaria Babiloni, Fabio Menegaz, Gloria Storti, Silvia F. Dept. of Computer Science University of Verona Verona Italy Dept. of Engineering for Innovation Medicine University of Verona Verona Italy Dept. of Molecular Medicine Sapienza University of Rome Rome Italy

In the context of Industry 5.0 and human-robot interaction, ensuring the safety of operators by avoiding human errors is crucial. Monitoring vigilance decrement is an essential aspect of this effort, aimed at mitigating safety risks and enhancing productivity. A potentially promising solution to this challenge is using a passive brain-computer interface (BCI) based on electroencephalography (EEG) recordings. However, its application in industrial settings has yet to be explored in-dept.. This study uses EEG data to introduce a novel experimental protocol and analysis pipeline to predict vigilance degradation in an industrial research laboratory. The dataset was gathered from ten healthy volunteers who observed a robotic arm for 23 min. The EEG power spectrum over time was computed using the continuous wavelet transform (CWT). After confirming growth in power for the α band using a linear regression model, we forecast its trend using four models. As a conventional approach, we used the vector autoregressive (VAR) model, serving as a reference for comparison with three deep learning architectures: a temporal convolutional network (TCN), a gated recurrent unit (GRU) and an encoder-decoder (ED)-GRU. The proposed ED-GRU model outperformed the others showing accurate forecasts (mean absolute error = 0.048, R² = 0.726) up to 5.5 s. The findings suggest that monitoring vigilance degradation in Industry 5.0 is a feasible strategy to prevent human accidents and reduced performance during repetitive tasks. © 2025 Nalecz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of sciences

关键词： Brain-computer interface Deep-learning Electroencephalography Forecasting Industry 5.0. Vigilance

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：