检索结果-内蒙古大学图书馆

Attention-Enhanced Hybrid Feature Aggregation Network for 3D Brain Tumor Segmentation 1

Challenge on Brain Tumor Segmentation, BraTS 2023, International Challenge on Cross-Modality Domain Adaptation for Medical Image Segmentation, CrossMoDA 2023, held in conjunction with the Medical Image Computing for computer Assisted Intervention Conference, MICCAI 2023

作者： Yazıcı, Ziya Ata Öksüz, İlkay Ekenel, Hazım Kemal Department of Computer Engineering Istanbul Technical University Istanbul Turkey Department of Computer Science and Engineering Qatar University Doha Qatar

ISBN: (数字)9783031761638

ISBN: (纸本)9783031761621

Glioblastoma is a highly aggressive and malignant brain tumor type that requires early diagnosis and prompt intervention. Due to its heterogeneity in appearance, developing automated detection approaches is challenging. To address this challenge, Artificial Intelligence (AI)-driven approaches in healthcare have generated interest in efficiently diagnosing and evaluating brain tumors. The Brain Tumor Segmentation Challenge (BraTS) is a platform for developing and assessing automated techniques for tumor analysis using high-quality, clinically acquired MRI data. In our approach, we utilized a multi-scale, attention-guided and hybrid U-Net-shaped model – GLIMS – to perform 3D brain tumor segmentation in three regions: Enhancing Tumor (ET), Tumor Core (TC), and Whole Tumor (WT). The multi-scale feature extraction provides better contextual feature aggregation in high resolutions and the Swin Transformer blocks improve the global feature extraction at deeper levels of the model. The segmentation mask generation in the decoder branch is guided by the attention-refined features gathered from the encoder branch to enhance the important attributes. Moreover, hierarchical supervision is used to train the model efficiently. Our model’s performance on the validation set resulted in 92.19, 87.75, and 83.18 Dice Scores and 89.09, 84.67, and 82.15 Lesion-wise Dice Scores in WT, TC, and ET, respectively. The code is publicly available at https://***/yaziciz/GLIMS. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Impact of Grid Strength on Sub-Synchronous Oscillations in AC Systems with Type-4 Wind Farms Integrated 50

Impact of Grid Strength on Sub-Synchronous Oscillations in A...

引用

50th Annual Conference of the IEEE Industrial Electronics Society, IECON 2024

作者： Zamzam, Tassneem Umar, Muhammad F. Qiblawey, Yazan Alassi, Abdulrahman Ghrayeb, Ali Abu-Rub, Haitham Department of Electrical and Computer Engineering Texas A&M University College Station United States Department of Electrical and Computer Engineering Texas A&M University at Qatar Doha Qatar Iberdrola Innovation Middle East Doha Qatar Department of Electrical and Computer Engineering College of Science and Engineering Hamad Bin Khalifa University Qatar Foundation Doha Qatar

ISBN: (纸本)9781665464543

The modern power system dominated by renewable energy resources such as Type-4 wind farms (WF) has recently seen significant increase in cases of sustained oscillations at the sub-synchronous frequency range that is typically referred to sub-synchronous oscillations (SSOs) in the existing literature. Although thorough studies have been conducted on prior types of WFs to understand the triggering factors for SSO, however, in the case of Type-4 WFs the triggering factors and causes still remain unclear. Therefore, this paper studies the SSO characteristics exhibited by Type-4 WFs operating within a weak grid. To emulate real-world grid conditions, an extensive AC network modeled after the IEEE 39-bus system is employed. Through simulations incorporating the integration of Type-4 WFs at various locations within the grid, the resulting effects on SSO triggering factors are analyzed. The investigation is structured around three distinct case studies, conducted at buses 31, 37, and 38. These studies involve varying network reactance to explore and establish the relationship between SSO events and the short-circuit ratio (SCR) of the network. The consistent findings across diverse case studies underscore the general relationship between SCR, grid strength, and SSOs. The analysis reaffirms the significant impact of increased reactance on SSO characteristics and confirms that weaker grids are more susceptible to SSOs. © 2024 IEEE.

关键词： Wind power

来源：评论

学校读者我要写书评

暂无评论

Identification of Medicinal Plants and Related Fungal Diseases based on Deep Learning Model 6

Identification of Medicinal Plants and Related Fungal Diseas...

引用

6th International Conference on Mobile Computing and Sustainable Informatics, ICMCSI 2025

作者： Kani, Mathina M. Mahalakshmi, R. Shunmugavel, V. Kareem, Abdul M. Department of Computer Science and Engineering Sethu Institute of Technology Virudhunagar India Department of Electrical and Electronics Engineering Quantanics TechServ pvt Ltd Madurai India

ISBN: (纸本)9798331522667

In today’s technological landscape, manual methods are increasingly being replaced by automated computer systems, streamlining tasks and enhancing efficiency. Within the realm of agricultural research, two critical areas plant identification and disease classification have shifted toward computerized solutions. With millions of plant species globally, medicinal plants hold particular significance due to their healing properties in traditional medicine. These readily accessible and culturally important remedies not only contribute to affordable healthcare but also sustain livelihoods in indigenous and rural communities. To address these challenges, our study evaluates the performance of two widely used pre-trained deep learning models U-Net and F-Net. By comparing their accuracy levels, we aim to enhance disease detection. Additionally, we present performance metrics such as accuracy and error rates, visualized through graphs. Our user interface (UI) leverages Kivy, an open-source multiplatform GUI development library for Python. Kivy enables cross-device app development, emphasizing code reusability and rapid prototyping, making it a valuable tool for interaction design across iOS, Android, Windows, OS X, and GNU/Linux platforms. © 2025 IEEE.

关键词： Medicinal chemistry

来源：评论

学校读者我要写书评

暂无评论

GestureSense: A Deep Learning-based Gesture Language Translator using VGG1 5

GestureSense: A Deep Learning-based Gesture Language Transla...

引用

5th International Conference on Image Processing and Capsule Networks, ICIPCN 2024

作者： Vyshnavi, Sure Lakshmi Chandana, Narra Ramya, Nallagorla Naga Sai Suvarna, B. Department of Computer Science and Engineering AP Guntur India

ISBN: (纸本)9798350367171

This research study introduces a Gesture Language Translator system leveraging state-of-the-art Convolutional Neural Network (CNN) architectures, including VGG-16 and ResNet. Through extensive experimentation, the efficacy of these models in capturing the complex features of sign language gestures is evaluated in this research study. To improve translation robustness and accuracy, the Sign Language Translator system combines the capabilities of the VGG-16 and ResNet-50 models using ensemble learning approaches. Furthermore, the project incorporates data preprocessing techniques to enhance the quality and consistency of input sign language images, ensuring optimal performance of the CNN models. Transfer learning strategies are also explored to adapt pre-trained CNN models to the task of sign language recognition, thereby reducing the need for extensive labeled data. Evaluation of the Sign Language Translator system is conducted on standard benchmark dataset ASL Alphabet demonstrating promising results in real-world scenarios. Through comprehensive experimentation and analysis, this research study promotes diversity and accessibility in communication while advancing assistive technologies for people with hearing impairments. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

High-Traversability and Precise Navigation for Mobile Robots in Constrained Environments

引用

IEEE Sensors Journal 2025年第12期25卷 22815-22826页

作者： Zhang, Muhua Ma, Lei Wu, Ying Shen, Kai Sun, Yongkui Leung, Henry Southwest Jiaotong University School of Electrical Engineering Chengdu611756 China University of Calgary Department of Electrical and Computer Engineering CalgaryABT2N 1N4 Canada

This paper presents an integrated sensing-planning navigation system for mobile robots used in industrial inspection. The system can be directly integrated with real robots. It enables the robot to pass through tight spaces, circumvent unexpected obstacles, and accurately reach the goal pose. This improves the robot’s adaptability to inspection tasks. For obstacle sensing, the system includes an obstacle parameterizer. Static obstacles in global grid map occupancy clusters are convexified. To improve the representation accuracy of the tight space, continuous and angled obstacles are automatically segmented into multiple convex polygons. This avoids inappropriate convex shapes. Unexpected obstacles in local grid maps are first represented by minimum bounding circles (MBCs) to simplify data association. They are then converted into inscribed convex polygons to ensure consistency with static obstacle descriptions. For motion planning, the system utilizes a two-stage planner operating on obstacle convex sets. When the robot is distant from the goal, the nonlinear model predictive control (NMPC) is constrained by inter-convex set discrete-time control barrier functions (DCBF-Convex), termed NMPC-DCBF-Convex, enabling effective navigation in tight spaces and around unexpected obstacles. As the robot nears the goal pose, DCBF-Convex conditions prune the sampling space of dynamic window approach (DWA), termed DWA-DCBF-Convex. It ensures high-frequency control and enables safe and precise arrival at the goal in constrained environments. Experiments in constrained scenarios validate the system’s real-world effectiveness. An inspection robot weighing over 70kg completes challenging navigation tasks at a maximum velocity of 0.5m/s, achieving an empirical terminal positional error of ±0.017m. © 2001-2012 IEEE.

关键词： Mobile robots

来源：评论

学校读者我要写书评

暂无评论

An effective abstract text summarisation using shark smell optimised bidirectional encoder representations from transformer

引用

International Journal of Business Intelligence and Data Mining 2023年第1期23卷 50-72页

作者： Nafees Muneera, M. Sriramya, P. Department of Computer Science and Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Tamilnadu Chennai602105 India Department of Computer Science and Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Chennai India

Recently, a vast amount of text data has increased rapidly and therefore information must be summarised to retrieve useful knowledge. First, the preprocessing module utilises the fixed-length stemming method, and then the segmentation module makes use of a pre-trained bidirectional encoder representations from transformers (BERT). The text of input is segmented with the utilisation of feedforward and multi-head attention layer. This BERT segmentation paradigm is adjoined alongside shark smell optimisation (SSO) methodology, and thus, the phrases that are extricated are employed to prepare the document stage of a dataset of Amazon merchandise assessment. This study aspires to create a concise summary and invigorating headlines, which grab the focus of the readers. This paper demonstrates that it performs by amalgamating the duo extractive and abstractive procedures employing a pipelined technique for creating a succinct summary that is later utilised for headline creation. Experimentation was executed on publicly accessible datasets – CNN/Daily Mail. Copyright © 2023 Inderscience Enterprises Ltd.

关键词： Signal encoding

来源：评论

学校读者我要写书评

暂无评论

DocTTT: Test-Time Training for Handwritten Document Recognition Using Meta-Auxiliary Learning

DocTTT: Test-Time Training for Handwritten Document Recognit...

引用

2025 IEEE/CVF Winter Conference on Applications of computer Vision, WACV 2025

作者： Gu, Wenhao Gu, Li Wang, Ziqiang Suen, Ching Yee Wang, Yang Concordia University Department of Computer Science and Software Engineering Canada

ISBN: (纸本)9798331510831

Despite recent significant advancements in Handwritten Document Recognition (HDR), the efficient and accurate recognition of text against complex backgrounds, diverse handwriting styles, and varying document layouts remains a practical challenge. Moreover, this issue is seldom addressed in academic research, particularly in scenarios with minimal annotated data available. In this paper, we introduce the DocTTT framework to address these challenges. The key innovation of our approach is that it uses test-time training to adapt the model to each specific input during testing. We propose a novel Meta-Auxiliary learning approach that combines Meta-learning and self-supervised Masked Autoencoder (MAE). During testing, we adapt the visual representation parameters using a self-supervised MAE loss. During training, we learn the model parameters using a meta-learning framework, so that the model parameters are learned to adapt to a new input effectively. Experimental results show that our proposed method significantly outperforms existing state-of-the-art approaches on benchmark datasets. © 2025 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

A Secure and Efficient Cloud Storage System Using Advanced Encryption Standard Algorithm for Data Protection 5th

A Secure and Efficient Cloud Storage System Using Advanced E...

引用

5th EAI International Conference on Cognitive Computing and Cyber Physical Systems, IC4S 2024

作者： Babu, Gurujukota Ramesh Chintalapati, Phaneendra Varma Sree, Pokkuluri Kiran Satish Kumar, Kode Asha, A.V.S. Maneesha, Boddu Department of Computer Science and Engineering Shri Vishnu Engineering College for Women Bhimavaram India

ISBN: (纸本)9783031770777

The Cloud Computing (CC) is a model which treats the resources as an integrated entity on the internet, cloud. Cloud computing is an unique environment or network in which process, access and maintenance are done by anywhere from the world. The CC is a customized internet-based web server. Data security is receiving more and more attention with the expanding usage of cloud storage systems and due to its challenging circumstances. Any cloud data storage model’s fundamental objective is to provide simple data access without compromising the security of the data. The main concerns of any cloud storage system are ensuring safety and security. Big data security and privacy have emerged as a problem that prevents the usage of cloud services by the organization. Earlier privacy-preserving methods had a number of drawbacks includes completely dependence on third parties, a lack of performance efficiency, accurate data analysis, and data privacy. This analysis presents a secure and efficient cloud storage system that uses the advanced encryption standard algorithm for data protection. The user files are encrypted with the AES (Advanced Encryption Standard) standard. The efficiency of the method that is being provided is evaluated in terms of execution, decryption, and encryption times. © ICST Institute for computer sciences, Social Informatics and Telecommunications engineering 2025.

关键词： Cloud security

来源：评论

学校读者我要写书评

暂无评论

VidChain: Chain-of-Tasks with Metric-based Direct Preference Optimization for Dense Video Captioning 39

VidChain: Chain-of-Tasks with Metric-based Direct Preference...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Lee, Ji Soo Kim, Jongha Na, Jeehye Park, Jinyoung Kim, Hyunwoo J. Department of Computer Science and Engineering Korea University Korea Republic of

ISBN: (纸本)157735897X

Despite the advancements of Video Large Language Models (VideoLLMs) in various tasks, they struggle with fine-grained temporal understanding, such as Dense Video Captioning (DVC). DVC is a complicated task of describing all events within a video while also temporally localizing them, which integrates multiple fine-grained tasks, including video segmentation, video captioning, and temporal video grounding. Previous VideoLLMs attempt to solve DVC in a single step, failing to utilize their reasoning capability. Moreover, previous training objectives for VideoLLMs do not fully reflect the evaluation metrics, therefore not providing supervision directly aligned to target tasks. To address such a problem, we propose a novel framework named VidChain comprised of Chain-of-Tasks (CoTasks) and Metric-based Direct Preference Optimization (M-DPO). CoTasks decompose a complex task into a sequence of sub-tasks, allowing VideoLLMs to leverage their reasoning capabilities more effectively. M-DPO aligns a VideoLLM with evaluation metrics, providing fine-grained supervision to each task that is well-aligned with metrics. Applied to two different VideoLLMs, VidChain consistently improves their fine-grained video understanding, thereby outperforming previous VideoLLMs on two different DVC benchmarks and also on the temporal video grounding task. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Visual Robustness Benchmark for Visual Question Answering (VQA)

Visual Robustness Benchmark for Visual Question Answering (V...

引用

2025 IEEE/CVF Winter Conference on Applications of computer Vision, WACV 2025

作者： Ishmam, Md Farhan Tashdeed, Ishmam Saadat, Talukder Asir Ashmafee, Md Hamjajul Mostofa Kamal, Abu Raihan Hossain, Md. Azam Islamic University of Technology Department of Computer Science and Engineering Bangladesh

ISBN: (纸本)9798331510831

Can Visual Question Answering (VQA) systems maintain their performance when deployed in the real world? Or are they susceptible to realistic corruption effects, e.g., image blur, which can be detrimental in sensitive applications such as medical VQA? While linguistic robustness has been thoroughly explored within the VQA literature, there has yet to be any significant work on visual robustness. In this work, we present the first large-scale benchmark to evaluate the visual robustness of VQA models, including multimodal large language models and zero-shot evaluation, and assess the strength of the realistic corruption effects. Additionally, we have designed several robustness evaluation metrics that quantify an aspect of robustness. These sub-metrics can be aggregated into a unified metric and tailored to fit a variety of use cases. The experiments reveal important insights into the relationships between model size, accuracy, and robustness against the visual corruptions. Our benchmark highlights the need for a balanced approach in model development that considers model performance without compromising robustness. © 2025 IEEE.

关键词： common corruptions corruption robustness evaluation metric multimodal multimodal benchmark multimodal evaluation multimodal learning multimodal robustness robustness robustness benchmark robustness evaluation robustness metric vision language vision language evaluation vision language metric vision language robustness visual corruption visual question answering visual robustness vqa vqa benchmark vqa evaluation vqa metric vqa robustness

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：