检索结果-内蒙古大学图书馆

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Liang Zhao Yu Shao Zhanxin Gang Chaoran Jia Yihan Chen Yi Yang School of Software Technology Dalian University of Technology Dalian China School of Reliability and Systems Engineering Beihang University Beijing China

Soft tissues sarcomas include striated muscle, fibrous tissue, fat, and other soft tissues. Simultaneously, their mortality rates are comparable to those of esophageal cancer, cervical cancer, and other cancers. Prior to surgical resection of patients, it is frequently necessary to study and diagnose the sarcoma area using MRI images in order to design a better surgical plan. However, artificial approaches for diagnosing the sarcoma region are time-consuming and error-prone. While the advent of artificial intelligence allows for computer-assisted diagnosis of the sarcoma region. Nevertheless, there is currently a scarcity of high-quality soft tissue sarcoma imaging data sets in relevant sectors. Thus, with the aim to investigate how to use multi-modal MRI images of patients with soft tissue sarcomas to segment the sarcoma area, we collect and process 15372 multi-modal MRI images in coronal of 40 patients with soft tissue sarcomas found in the thigh, which we subsequently combine with the help of several clinicians to mark the sarcoma area. The multi-modal MRI imaging data set of soft tissue sarcoma is therefore acquired by a number of preprocessing techniques. The sarcoma area is then segmented using a multi-encoder and single-decoder network that adapts to multiple input modalities. For motivating the network to learn the important semantic features of different modalities, we design a feature fusion strategy mechanism that is applied to the skip connection. Additionally, self-supervised learning is being investigated to address the issue of a small number of data points in the data set. Experiments show that our network can achieve the highest Dice score of 57.76% on our data set. Our code and the dataset are available at https://***/syaxx0819/The-Multimodal-Soft-Tissue-Sarcoma-Image-Dataset.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Optimality Conditions for Model Predictive Control: Rethinking Predictive Model Design

arXiv

引用

arXiv 2024年

作者： Anand, Akhil S. Kordabad, Arash Bahari Zanon, Mario Gros, Sebastien Norwegian University of Science and Technology Trondheim7491 Norway Max Planck Institute for Software Systems Kaiserslautern Germany IMT School for Advanced Studies Lucca55100 Italy

Optimality is a critical aspect of Model Predictive Control (MPC), especially in economic MPC. However, achieving optimality in MPC presents significant challenges, and may even be impossible, due to inherent inaccuracies in the predictive models. Predictive models often fail to accurately capture the true system dynamics, such as in the presence of stochasticity, leading to suboptimal MPC policies. In this paper, we establish the necessary and sufficient conditions on the underlying prediction model for an MPC scheme to achieve closed-loop optimality. Interestingly, these conditions are counterintuitive to the traditional approach of building predictive models that best fit the data. These conditions present a mathematical foundation for constructing models that are directly linked to the performance of the resulting MPC scheme. © 2024, CC BY.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Discussion Manipulation, Language and Domain Dependent Models: An Overview

Discussion Manipulation, Language and Domain Dependent Model...

引用

Zooming Innovation in Consumer Electronics International Conference (ZINC)

作者： Filip Chalás Igor Stupavský Valentino Vranić Institute of Informatics Information Systems and Software Engineering Faculty of Informatics and Information Technologies Slovak University of Technology in Bratislava

The amount of antisocial online behavior (AOB) in the political field has been on the rise in recent years. In the research, we are dealing with the reason for confusing AOB with other forms of antisocial behavior. We specifically dealt with NLP in the political field in the Slovak language, with a specific number of prefixes and suffixes. In this paper, we focus on the area of detection of manipulation of political discussions. The data source consisted of previously collected data in the period from April 2018 with an update until November 19, 2022 realized within this research from the site ***, which we supplemented with new claims. We trained several classification models on preprocessed data and evaluated them. For multi-class classification, the best results were achieved using logistic regression and a support vector machine trained on the resampled dataset—both achieving an accuracy of 0.56 and a macro F1 score of 0.39. In the case of binary classification, the best results were achieved by logistic regression—accuracy 0.7 and macro F1 score 0.56. These models could help detect manipulation in online political discussions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multimodal contrastive learning with neuroimaging and cognitive tests for Alzheimer’s disease diagnosis

Multimodal contrastive learning with neuroimaging and cognit...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Liang Zhao Jian Zhang Bo Xu Yi Yang Yangqianhui Zhang Ruixin Ma School of Software Technology Dalian University of Technology Dalian China School of Reliability and Systems Engineering Beihang University Beijing China School of Mechanical Engineering Zhejiang University Hangzhou China

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

Alzheimer’s disease (AD) is a neurological illness that causes cognitive impairment. Computer-aided diagnosis can help diagnose Alzheimer’s disease early before clinical symptoms appear. Currently, many deep learning methods show good performance in AD diagnosis. Still, most of these methods are based on single/multimodal neuroimaging, leading to a one-sided approach to disease modelling. Combining neuroimaging, cognitive tests, and demographics can significantly improve model performance and reduce the negative impact of noise. This study proposes a multimodal model that introduces contrastive learning, extracting and fusing feature representations separately from cognitive tests and neuroimaging data. After that, contrastive learning based on similarity is employed for both modalities’ features, assisting the network in learning cross-modal features. Moreover, the hybrid attention mechanism of the Transformer encoder is explored for feature fusion. Experimental results on 2082 cases from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset validate the effectiveness of our proposed multimodal model.

关键词： Neuroimaging Attention mechanisms Computational modeling Noise Contrastive learning Predictive models Feature extraction Transformers Alzheimer's disease Medical diagnostic imaging

来源：评论

学校读者我要写书评

暂无评论

Stereo Image Super-Resolution via Disparity Estimation and Domain Diffusion 9th

Stereo Image Super-Resolution via Disparity Estimation and D...

引用

9th International Symposium on Artificial Intelligence and Robotics, ISAIR 2024

作者： Wang, Wanjun Ma, Chunyan Zhu, Hongjun Xu, Kai Han, Huihui School of Computer and Software Engineering Anhui Institute of Information Technology WuHu241199 China School of Artificial Intelligence Department of Information Hohai University NanJing210024 China Platforms of Automotive Intelligent Software and Systems Anhui Institute of Information Technology WuHu241199 China

ISBN: (纸本)9789819629107

Image super-resolution (SR) and disparity estimation are closely related in stereo images, with effective utilization of disparity maps enhancing SR performance. This paper proposes a stereo image super-resolution reconstruction network utilizing disparity estimation and a domain diffusion model. The network integrates disparity estimation with diffusion models for feature summation and domain space construction, enabling interaction between forward and backward diffusion memory units. To further improve reconstructed image quality, we introduce a novel memory unit inspired by long short-term memory concepts. Disparity loss and structural similarity are included in the loss function to achieve more accurate results. Experiments show that our method outperforms existing techniques. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Stereo image processing

来源：评论

学校读者我要写书评

暂无评论

Effectiveness of Using Deep Learning for Blister Blight Identification in Sri Lankan Tea

Effectiveness of Using Deep Learning for Blister Blight Iden...

引用

2023 International Research Conference on Smart Computing and systems engineering, SCSE 2023

作者： Hewawitharana, G.H.A.U. Nawarathne, U.M.L.A. Hassan, A.S.F. Wijerathna, Lochana M. Sinniah, G.D. Vidhanaarachchi, Samitha P. Wickramarathne, Jagath Wijekoon, Janaka L. Sri Lanka Institute of Information Technology Department of Information Technology Malabe Sri Lanka Tea Research Institute of Sri Lanka Plant Pathology Division Talawakelle Sri Lanka Sri Lanka Institute of Information Technology Department of Computer Science and Software Engineering Malabe Sri Lanka Military Technological College Systems Engineering Department Muscat Oman

ISBN: (纸本)9798350341454

Ceylon tea industry faces a major challenge in the form of pathogen-induced crop loss, with Blister Blight (BB) caused by Exobasidium vexans posing the greatest threat, leading to harvest losses of over 30%. This fungus attacks the tender tea shoots, resulting in a direct negative impact on the tea harvest. This paper presents a system to identify the suspicious tea leaves and BB disease at its early stages along with an assessment of severity, offering a potential solution to this critical issue. By utilizing real-time object detection, the system filters out non-tea leaves from the captured initial image of a segment of a tea plant. The identified tea leaves are then subjected to BB identification and severity assessment based on differing visual symptoms of the BB stages. This approach enables the system to accurately identify BB in the initial stage and severity stage, allowing for timely and targeted intervention to minimize crop losses. The YOLOv8 model has been able to correctly identify 98% of the objects it has detected as relevant (precision), and it has been able to correctly identify 96% of all the relevant objects present in the scene (recall). The Residual Network 50 (Resnet50) convolutional neural network (CNN) model was selected as the final model, achieving an accuracy of 89.90% during the training phase and an accuracy of 88.26% during the testing phase. © 2023 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

M2IST: Multi-Modal Interactive Side-Tuning for Efficient Referring Expression Comprehension

引用

IEEE Transactions on Circuits and systems for Video technology 2025年

作者： Liu, Xuyang Liu, Ting Huang, Siteng Xin, Yi Hu, Yue Qin, Long Wang, Donglin Wu, Yuanyuan Chen, Honggang Sichuan University College of Electronics and Information Engineering Chengdu610065 China National University of Defense Technology College of Systems Engineering Changsha410072 China Westlake University School of Engineering Hangzhou310030 China Nanjing University State Key Laboratory for Novel Software Technology Nanjing210000 China Chengdu610059 China Yunnan University Yunnan Key Laboratory of Software Engineering Kunming650600 China

Referring expression comprehension (REC) is a vision-language task to locate a target object in an image based on a language expression. Fully fine-tuning general-purpose pre-trained vision-language foundation models for REC yields impressive performance but becomes increasingly costly. Parameter-efficient transfer learning (PETL) methods have shown strong performance with fewer tunable parameters. However, directly applying PETL to REC faces two challenges: (1) insufficient multi-modal interaction between pre-trained vision-language foundation models, and (2) high GPU memory usage due to gradients passing through the heavy vision-language foundation models. To this end, we present M2IST: Multi-Modal Interactive Side-Tuning with M3ISAs: Mixture of Multi-Modal Interactive Side-Adapters. During fine-tuning, we fix the pre-trained uni-modal encoders and update M3ISAs to enable efficient vision-language alignment for REC. Empirical results reveal that M2IST achieves better performance-efficiency trade-off than full fine-tuning and other PETL methods, requiring only 2.11% tunable parameters, 39.61% GPU memory, and 63.46% training time while maintaining competitive performance. Our code is released at https://***/xuyang-liu16/M2IST. © 1991-2012 IEEE.

关键词： Visual languages

来源：评论

学校读者我要写书评

暂无评论

Materials Behaviour Analysis of 3D Printed Brass-PLA Filament 7

Materials Behaviour Analysis of 3D Printed Brass-PLA Filamen...

引用

7th International Conference on Mechanical engineering Research 2023, ICMER 2023

作者： Rajan, Dinesh Samykano, M. Suraparaju, S.K. Moorthy, K. Kadirgama, K. Ramasamy, D. Pandey, A.K. Faculty of Mechanical and Automotive Engineering Technology Universiti Malaysia Pahang Al-Sultan Abdullah Pekan Pahang26600 Malaysia Centre for Research in Advanced Fluid and Process University Malaysia Pahang Al-Sultan Abdullah Lebuhraya Tun Razak Pahang Kuantan Gambang26300 Malaysia Faculty of Computer Systems & Software Engineering Universiti Malaysia Pahang Al-Sultan Abdullah Kuantan Malaysia School of Engineering and Technology Sunway University No. 5 Jalan Universiti Bandar Sunway Selangor Darul Ehsan Petaling Jaya47500 Malaysia

The current study aims to bridge a crucial gap in existing research, potentially paving the way for a groundbreaking transformation in the development and application of PLA/Brass composites within diverse industries such as aerospace, automotive, consumer goods, and medical devices. The primary objective of this research is to assess the mechanical properties of a composite material made up of Polylactic Acid (PLA) and Brass, produced using Fused Deposition Modelling (FDM) 3D printing technology. Brass, renowned for its exceptional mechanical properties, has been integrated into PLA to form this composite material. The study employs various analytical techniques, including Fourier Transform Infrared Spectroscopy (FTIR), Scanning Electron Microscopy (SEM), and Energy-Dispersive X-ray Spectroscopy (EDX), to scrutinize the chemical and physical characteristics of the PLA/Brass composite. This research revolves around exploring the impact of different printing parameters on the mechanical behavior of the printed specimens. The investigation delves into aspects such as tensile strength, compression resistance, bending properties, and impact resistance. To achieve this, test specimens with varying compositions have been produced using a Raise3D N2 Plus FDM 3D printer, with careful manipulation of printing parameters such as layer height and printing speed. The compositional variations range from 15% wt. to 80% wt., with layer height values spanning 0.25 mm, 0.30 mm, and 0.35 mm, and printing speeds ranging from 20 mm/s to 40 mm/s. The outcomes of this research have revealed the distinct influences of specific printing parameters on various mechanical properties. For example, in the context of tensile testing, it was observed that the combination of a layer height of 0.25 mm and a printing speed of 30 mm/s resulted in the highest elastic modulus. Similarly, the study provides crucial insights into optimizing PLA/Brass composite material properties through controlled addit

关键词： Scanning electron microscopy

来源：评论

学校读者我要写书评

暂无评论

Identifying and Involving the Real End User in software Development: Towards a Pattern Language 22

Identifying and Involving the Real End User in Software Deve...

引用

Proceedings of the 27th European Conference on Pattern Languages of Programs

作者： Mohammad Daud Haiderzai Valentino Vranić Institute of Informatics Information Systems and Software EngineeringFaculty of Informatics and Information Technologies Slovak University of Technology Slovakia

ISBN: (纸本)9781450395946

software development demand and identification of end users involvement have been increasing rapidly, but identifying the real end user and involving them in software development is challenging for software developers reported in the literature. One of the problems is the difficulty of designing the business domain. It takes longer to find the actual business domain when different and inexpert domain end users are engaged with developers. Another problem is the difficulty of incorporating all types of end users needs in software development. Several expectations and different information make the development process more complex and overload the project scope. software development also takes more time, and the project goal is challenging. The above problems are the leading causes that create a complex process and difficulty in software development. To overcome these problems, we established a pattern language composed of eight organizational patterns observed and documented during in practical work on many projects. These patterns and other known organizational patterns relation are described in a patterns story of a real company.

关键词： software development user involvement Organizational patterns

来源：评论

学校读者我要写书评

暂无评论

Diversifying Message Aggregation in Multi-Agent Communication Via Normalized Tensor Nuclear Norm Regularization

Diversifying Message Aggregation in Multi-Agent Communicatio...

引用

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

作者： Yuanzhao Zhai Kele Xu Bo Ding Dawei Feng Zijian Gao Huaimin Wang National University of Defense Technology Changsha China Key Laboratory of Software Engineering for Complex Systems Changsha China

The use of graph attention networks (GAT) in communication-enhanced multi-agent reinforcement learning (Comm-MARL) has become prevalent. While successful, GAT can lead to homogeneity in the strategies of message aggregation, which can severely limit multi-agent coordination. To address this challenge, we study the adjacency tensor of the communication graph. Then we define a new nuclear tensor rank and its convex surrogate, the normalized tensor nuclear norm to measure the homogeneity of message aggregation. Leveraging the norm, we further propose a plug-and-play regularizer on the adjacency tensor, named Normalized Tensor Nuclear Norm Regularization (NTNNR), to actively enrich the diversity of message aggregation during the training stage. NTNNR is agnostic to specific Comm-MARL algorithms and can be flexibly integrated with different graph-attention methods. Empirical results demonstrate that aggregating messages using NTNNR-enhanced GAT can improve the efficiency of the training and achieve higher asymptotic performance than existing message aggregation methods.

关键词： Training Tensors Signal processing algorithms Reinforcement learning Signal processing Acoustics Speech processing

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：