检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Chen, Chaoqi Tang, Luyao Huang, Hui College of Computer Science and Software Engineering Shenzhen University China School of Informatics Xiamen University China

Since deep learning models are usually deployed in non-stationary environments, it is imperative to improve their robustness to out-of-distribution (OOD) data. A common approach to mitigate distribution shift is to regularize internal representations or predictors learned from in-distribution (ID) data to be domain invariant. Past studies have primarily learned pairwise invariances, ignoring the intrinsic structure and high-order dependencies of the data. Unlike machines, humans recognize objects by first dividing them into major components and then identifying the topological relation of these components. Motivated by this, we propose Reconstruct and Match (REMA), a general learning framework for object recognition tasks to endow deep models with the capability of capturing the topological homogeneity of objects without human prior knowledge or fine-grained annotations. To identify major components from objects, REMA introduces a selective slot-based reconstruction module to dynamically map dense pixels into a sparse and discrete set of slot vectors in an unsupervised manner. Then, to model high-order dependencies among these components, we propose a hypergraph-based relational reasoning module that models the intricate relations of nodes (slots) with structural constraints. Experiments on standard benchmarks show that REMA outperforms state-of-the-art methods in OOD generalization and test-time adaptation settings. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A proposed brain tumor detection algorithm using Multi wavelet Transform (MWT)

引用

Materials Today: Proceedings 2022年 65卷 2731-2737页

作者： Mahmoud Al-Jawher, Waleed A. Awad, Sarah H. Uruk University Department of computer engineering Iraq Iraqi Commission for computers and informatics Iraq

Magnetic resonance imaging)MRI) is a technological development in the medical field. It is used to give images of the human body with high accuracy and good quality, facilitating the process of identifying and classifying diseases in the human body. One of the diseases that are diagnosed using MRI is a brain tumor, MRI which helps in the early diagnosis of the tumor, and helps the doctor to diagnose and identify the tumor and thus make the fastest medical decision. This paper used to identification and extraction of brain tumors from magnetic resonance images based on Multi Wavelet Transform (MWT) and image processing techniques. This project Multi Wavelet Transform (MWT) used in preprocessing stage to enhance the input image and denoising. Thresholding based is used for segmentation methods, classification of brain tumor is done using Statistical Classification Methods to classify the input MRI of brain into normal or abnormal. Firstly the location of tumor was presented & next its area computed. © 2022

关键词： Magnetic resonance imaging

来源：评论

学校读者我要写书评

暂无评论

Novel Framework of Segmentation 3D MRI of Brain Tumors

引用

computers, Materials & Continua 2023年第2期74卷 3489-3502页

作者： Ibrahim Mahmoud El-Henawy Mostafa Elbaz Zainab H.Ali Noha Sakr Department of Computer Science Faculty of Computers and InformaticsZagazig UniversityZagazigEgypt Department of Computer Science Faculty of Computers and InformaticsKafrelsheikh UniversityKafrelsheikhEgypt Department of Embedded Network Sys.and Tec. Faculty of Artificial IntelligenceKafrelsheikh UniversityKafrelsheikhEgypt Computers Engineering and Control Systems Department Faculty of EngineeringMansoura UniversityMansouraEgypt

Medical image segmentation is a crucial process for computer-aided diagnosis and *** image segmentation refers to portioning the images into small,disjointed parts for simplifying the processes of analysis and *** and speckle noise are different types of noise in magnetic resonance imaging(MRI)that affect the accuracy of the segmentation process ***,image enhancement has a significant role in MRI *** paper proposes a novel framework that uses 3D MRI images from Kaggle and applies different diverse models to remove Rician and speckle noise using the best possible noise-free *** proposed techniques consider the values of Peak Signal to Noise Ratio(PSNR)and the level of noise as inputs to the attention-U-Net model for segmentation of the *** framework has been divided into three stages:removing speckle and Rician noise,the segmentation stage,and the feature extraction *** framework presents solutions for each problem at a different stage of the *** the first stage,the framework uses Vibrational Mode Decomposition(VMD)along with Block-matching and 3D filtering(Bm3D)algorithms to remove the ***,the most significant Rician noise-free images are passed to the three different methods:Deep Residual Network(DeRNet),Dilated Convolution Auto-encoder Denoising Network(Di-Conv-AE-Net),andDenoising Generative Adversarial Network(DGAN-Net)for removing the speckle *** Bm3D have achieved PSNR values for levels of noise(0,0.25,0.5,0.75)for reducing the Rician noise by(35.243,32.135,28.214,24.124)and(36.11,31.212,26.215,24.123)*** framework also achieved PSNR values for removing the speckle noise process for each level as follows:(34.146,30.313,28.125,24.001),(33.112,29.103,27.110,24.194),and(32.113,28.017,26.193,23.121)forDeRNet,Di-Conv-AE-Net,and DGAN-Net,*** experiments that have been conducted have proved the efficiency of the proposed framework a

关键词： MRI Rician noise speckle noise segmentation deep learning

来源：评论

学校读者我要写书评

暂无评论

Design and In-training Optimization of Binary Search ADC for Flexible Classifiers 25

Design and In-training Optimization of Binary Search ADC for...

引用

30th Asia and South Pacific Design Automation Conference, ASP-DAC 2025

作者： Lozano Duarte, Paula Carolina Afentaki, Florentia Zervakis, Georgios Tahoori, Mehdi Department of Computer Science Karlsruhe Inst. of Tech. Karlsruhe Germany Computer Engineering and Informatics Department Univ. of Patras Patras Greece

ISBN: (纸本)9798400706356

Flexible Electronics (FE) offer distinct advantages, including mechanical flexibility and low process temperatures, enabling extremely low-cost production. To address the demands of applications such as smart sensors and wearables, flexible devices must be small and operate at low supply voltages. Additionally, target applications often require classifiers to operate directly on analog sensory input, necessitating the use of Analog to Digital Converters (ADCs) to process the sensory data. However, ADCs present serious challenges, particularly in terms of high area and power consumption, especially when considering stringent area and energy budget. In this work, we target common classifiers in this domain such as MLPs and SVMs and present a holistic approach to mitigate the elevated overhead of analog to digital interfacing in FE. First, we propose a novel design for Binary Search ADC that reduces area overhead 2× compared with the state-of-the-art Binary design and up to 5.4× compared with Flash ADC. Next, we present an in-training ADC optimization in which we keep the bare-minimum representations required and simplifying ADCs by removing unnecessary components. Our in-training optimization further reduces on average the area in terms of transistor count of the required ADCs by 5× for less than 1% accuracy loss. © 2025 Institute of Electrical and Electronics Engineers Inc.. All rights reserved.

关键词： Budget control

来源：评论

学校读者我要写书评

暂无评论

Feature-preserving color pencil drawings from photographs

引用

Computational Visual Media 2023年第4期9卷 807-825页

作者： Dong Wang Guiqing Li Chengying Gao Shengwu Fu Yun Liang College of Mathematics and Informatics South China Agricultural UniversityGuangzhouChina School of Computer Science&Engineering South China University of TechnologyGuangzhouChina School of Data and Computer Science Sun Yat-sen UniversityGuangzhouChina

Color pencil drawing is well-loved due to its rich *** paper proposes an approach for generating feature-preserving color pencil drawings from *** mimic the tonal style of color pencil drawings,which are much lighter and have relatively lower saturation than photographs,we devise a lightness enhancement mapping and a saturation reduction *** lightness mapping is a monotonically decreasing derivative function,which not only increases lightness but also preserves input photograph *** saturation is usually related to lightness,so we suppress the saturation dependent on lightness to yield a harmonious ***,two extremum operators are provided to generate a foreground-aware outline map in which the colors of the generated contours and the foreground object are *** experiments show that color pencil drawings generated by our method surpass existing methods in tone capture and feature preservation.

关键词： non-photorealistic rendering pencil drawings image editing feature preservation

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Analysis of Histopathology Images for Breast Cancer Detection: A Comparative Study of ResNet and VGG Architectures 18

Deep Learning Analysis of Histopathology Images for Breast C...

引用

18th International Workshop on Semantic and Social Media Adaptation and Personalization, SMAP 2023

作者： Kanavos, Andreas Mylonas, Phivos Ionian University Department of Informatics Corfu Greece University of West Attica Department of Informatics and Computer Engineering Athens Greece

ISBN: (纸本)9798350327717

Medical image analysis has undergone significant advancements with the emergence of deep learning techniques, offering great promise in improving diagnostic precision and expediting patient care. This research investigates the effectiveness of ResNet and VGG architectures in detecting breast cancer through the analysis of histopathology images. By meticulously fine-Tuning hyperparameters and optimizers, we establish robust and accurate deep learning models. Our findings reveal that the ResNet model with the SGD optimizer excels, surpassing the performance of VGG in terms of accuracy and F1-score. However, employing transfer learning with pre-Trained VGG16 and ResNet50 networks does not yield competitive results, potentially due to disparities in input image size and data distribution. The primary focus of this study is to address the critical challenge of early breast cancer detection, ultimately leading to enhanced patient outcomes. By exploring state-of-The-Art deep learning architectures and methodologies, we contribute to the growing body of research aimed at leveraging artificial intelligence for medical diagnosis. © 2023 IEEE.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

ShipQuest: A Bottom-Up Approach for Building a Modern Greek Maritime Ontology 18

ShipQuest: A Bottom-Up Approach for Building a Modern Greek ...

引用

18th International Workshop on Semantic and Social Media Adaptation and Personalization, SMAP 2023

作者： Mathe, Eirini Mouratidis, Despoina Voutos, Yorghos Kermanidis, Katia Mylonas, Phivos Kanavos, Andreas Ionian University Department of Informatics Corfu Greece University of West Attica Department of Informatics and Computer Engineering Athens Greece

ISBN: (纸本)9798350327717

The ShipQuest Ontology is a pioneering legal maritime ontology designed for modeling and integrating legal information in the Greek language. This standardized representation of legal concepts within the maritime domain offers a structured approach to managing legal data. By formalizing legal information in Greek, the ShipQuest Ontology ensures semantic interoperability and seamless integration of diverse legal data sources. It encompasses essential classes, subclasses, properties, and relationships pertaining to legal aspects of maritime activities, including legal entities, contracts, regulations, liabilities, and more. The ShipQuest Ontology plays a critical role in stream-lining legal operations within the maritime industry, offering a unified and comprehensive knowledge base for legal practitioners, researchers, and decision-makers. Its adoption fosters collaboration and knowledge exchange within the maritime community, promoting standardized legal terminology and practices. This paper delves into the construction and utilization of the ShipQuest Ontology, detailing its architecture, development process, and implementation. Furthermore, we showcase its practical benefits by illustrating its application in real-world scenarios. By addressing the challenges of legal data heterogeneity and facilitating semantic consistency, the ShipQuest Ontology sets a precedent for enhanced efficiency and effectiveness in the legal domain of maritime activities. © 2023 IEEE.

关键词： Ontology

来源：评论

学校读者我要写书评

暂无评论

Factors Influencing Participation in Data Crowdsourcing: A Systematic Literature Review 8

Factors Influencing Participation in Data Crowdsourcing: A S...

引用

8th International Conference on informatics and Computing, ICIC 2023

作者： Panjaitan, Re. Miracle Setiawan, Budi Darma Brawijaya University Faculty of Computer Science Department of Informatics Engineering Malang Indonesia

ISBN: (纸本)9798350342604

Crowdsourcing is becoming a trending concept used by many organizations, without being limited to a particular field. As of 2018, crowdsourcing was used by 85% of the largest worldwide companies. This trend of crowdsourcing would be increasing as it offers numerous advantages to companies in achieving their goals, such as the capability of obtaining data and information from the crowd. The crowd-provided data and information contain the essence of human ingenuity, which enables companies to do problem-solving with a human-centered approach. The more contributors participate, the more likely the crowdsourcing program will succeed. Several studies give different lists of factors that influence the participation rate of the crowd. However, there's a lack of discussion about the factors and it leads to a situation where organizations need to examine many factors when developing a crowdsourcing platform to get great crowd engagement. Therefore, this study conducted a Systematic Literature Review (SLR) to perform screening and analysis of preceding research systematically. The factors in each of these literatures are generated by considering the case of the study or the region of the people related to the research. This diversity helps identify the factors influencing contributor participation in data crowdsourcing from research all over the world. A study reflecting on these research findings would help answer research questions thoroughly. This study found that there are 19 factors that positively influence contributor participation in data crowdsourcing out of 26 factors found in the 16 selected documents. © 2023 IEEE.

关键词： Crowdsourcing

来源：评论

学校读者我要写书评

暂无评论

Comparison of Single Exponential Smoothing And Holt-Winter Exponential Smoothing Methods in Sales Commercial Business 11

Comparison of Single Exponential Smoothing And Holt-Winter E...

引用

11th International Conference on Cyber and IT Service Management, CITSM 2023

作者： Arini Suseno, Hendra Bayu Maulana, Fahmi Nur Matin, Iik Muhamad Malik Uin Syarif Hidayatullah Jakarta Informatics Jakarta Indonesia Jakarta State Polytechnic Computer and Informatics Engineering Jakarta Indonesia

ISBN: (纸本)9798350305968

Sales is an important aspect of running a commercial business. To increase sales, it must be accompanied by adequate availability of goods so that consumer needs can be met. However, the challenge in availability is that it is difficult to predict future consumer needs, so overstock and shortages often occur. On the other hand, the organization must be able to provide goods on time. This can be detrimental to the organization. For this reason, many solutions have been developed to overcome this problem, namely by predicting the availability of goods in the future using the single exponential smoothing and Holt-winter exponential smoothing methods. The single exponential smoothing method was used by previous researchers to provide good predictions on annual historical data. However, by using the same data converted into a monthly format, trend and seasonal factors were discovered. Therefore, we perform ex-inventory predictions using a single exponential smoothing method and then compare their performance. Trend and seasonal factors can be handled with the Holt-Winter exponential smoothing method. To determine the best prediction results from both methods for monthly data, we used the MAPE (Mean Absolute Percentage Error) method. The research results show that the Holt-winter exponential smoothing method has better performance with a smaller error value compared to single exponential smoothing, namely 1.828%. © 2023 IEEE.

关键词： Sales

来源：评论

学校读者我要写书评

暂无评论

Towards Human-Like Machine Comprehension: Few-Shot Relational Learning in Visually-Rich Documents 30

Towards Human-Like Machine Comprehension: Few-Shot Relationa...

引用

Joint 30th International Conference on Computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024

作者： Wang, Hao Li, Tang Chu, Chenhui Zhu, Nengjun Wang, Rui Zhu, Pinpin School of Computer Engineering and Science Shanghai University China Graduate School of Informatics Kyoto University Japan Department of Computer Science and Engineering Shanghai Jiao Tong University China

ISBN: (纸本)9782493814104

Key-value relations are prevalent in Visually-Rich Documents (VRDs), often depicted in distinct spatial regions accompanied by specific color and font styles. These non-textual cues serve as important indicators that greatly enhance human comprehension and acquisition of such relation triplets. However, current document AI approaches often fail to consider this valuable prior information related to visual and spatial features, resulting in suboptimal performance, particularly when dealing with limited examples. To address this limitation, our research focuses on few-shot relational learning, specifically targeting the extraction of key-value relation triplets in VRDs. Given the absence of a suitable dataset for this task, we introduce two new few-shot benchmarks built upon existing supervised benchmark datasets. Furthermore, we propose a variational approach that incorporates relational 2D-spatial priors and prototypical rectification techniques. This approach aims to generate relation representations that are more aware of the spatial context and unseen relation in a manner similar to human perception. Experimental results demonstrate the effectiveness of our proposed method by showcasing its ability to outperform existing methods. This study also opens up new possibilities for practical applications. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

关键词： Variational techniques

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：