检索结果-内蒙古大学图书馆

6th International Conference on Medical Imaging with Deep Learning, MIDL 2023

作者： Abutalip, Kudaibergen Saeed, Numan Khan, Mustaqeem Saddik, Abdulmotaleb El Computer Vision Department Mohamed Bin Zayed University of Artificial Intelligence Abu Dhabi United Arab Emirates School of Electrical Engineering and Computer Science University of Ottawa Ottawa Canada

Variability in staining protocols, such as different slide preparation techniques, chemicals, and scanner configurations, can result in a diverse set of whole slide images (WSIs). This distribution shift can negatively impact the performance of deep learning models on unseen samples, presenting a significant challenge for developing new computational pathology applications. In this study, we propose a method for improving the generalizability of convolutional neural networks (CNNs) to stain changes in a single-source setting for semantic segmentation. Recent studies indicate that style features mainly exist as covariances in earlier network layers. We design a channel attention mechanism based on these findings that detects stain-specific features and modify the previously proposed stain-invariant training scheme. We reweigh the outputs of earlier layers and pass them to the stain-adversarial training branch. We evaluate our method on multi-center, multi-stain datasets and demonstrate its effectiveness through interpretability analysis. Our approach achieves substantial improvements over baselines and competitive performance compared to other methods, as measured by various evaluation metrics. We also show that combining our method with stain augmentation leads to mutually beneficial results and outperforms other techniques. Overall, our study makes significant contributions to the field of computational pathology. © 2023 CC-BY 4.0, K. Abutalip, N. Saeed, M. Khan & A.E. Saddik.

关键词： Network layers

来源：评论

学校读者我要写书评

暂无评论

Autonomous Unmanned Aerial Vehicle Indoor Navigation Based on Deep Learning Approaches 11

Autonomous Unmanned Aerial Vehicle Indoor Navigation Based o...

引用

11th RSI International Conference on Robotics and Mechatronics, ICRoM 2023

作者： Ghasemi, Hamed Balazadeh, Elnaz Malekpour, Shayan Kalhor, Ahmad Masouleh, Mehdi Tale University of Tehran Human and Robot Interaction Laboratory School of Electrical and Computer Engineering Tehran Iran University of Tehran Human and Robot Interaction Laboratory Tehran Iran University of Tehran School of Electrical and Computer Engineering Department of Electrical and Computer Engineering Tehran Iran

ISBN: (纸本)9798350308105

This article presents two approaches based on deep learning for achieving autonomous navigation and subsequently, conducts an in-depth comparative analysis to elaborate their respective performance. Both methods leverage neural networks to facilitate navigation. The first approach employs a deep convolutional neural network to identify suitable areas for forward movement, while the second method utilizes the drone's front camera to capture images, which are then processed by a generative network to generate corresponding depth images. Importantly, both approaches rely solely on an RGB camera, resulting in cost savings for the robot and reduced energy consumption. The training data used for classifying collision-free and appropriate paths were gathered from simulated environments closely resembling real indoor settings. The proposed model calculates the probability associated with each segment of the input image, enabling the selection of the most viable area for drone flight. Incorporating generative networks as a depth estimation module enhances flight accuracy, making navigation in uncharted environments more straightforward. The accuracy of depth estimation was evaluated using standard criteria, achieving an accuracy rate of 84%. This result underscores the effectiveness of the proposed method. © 2023 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Classification of Lung Nodule from CT and PET/CT Images Using Artificial Neural Network

Classification of Lung Nodule from CT and PET/CT Images Usin...

引用

International Conference on Machine vision and Augmented Intelligence, MAI 2022

作者： Hansdah, Malho Singh, Koushlendra Kumar Machine Vision and Intelligence Lab Department of Computer Science and Engineering National Institute of Technology Jamshedpur Jharkhand Jamshedpur831014 India

ISBN: (纸本)9789819901883

This work aims to design and develop an artificial neural network (ANN) architecture for the classification of cancerous tissue in the lung. A sequential model is used for the machine learning process. ReLU and Sigmoid activation functions have been used to supply weights to the model. The present work encompasses detecting and classifying the tumor cells into four categories. The four types of lung cancer nodules are adenocarcinoma, squamous-cell carcinoma, large-cell carcinoma, and small-cell carcinoma. Computed tomography (CT) and Positron emission tomography (PET) scan DICOM images are used for the classification. The proposed approach has been validated with the subset of the original dataset. A total of 6500 images have been taken in the experiment. The approach is to feed the CT scan images into ANNs and classify the image as the correct type. The dataset is provided by The Cancer Imaging Archive (TCIA). The dataset is titled "A Large-Scale CT and PET/CT Dataset for Lung Cancer Diagnosis." The tumor cells are classified using the ANN architecture with 99.6% of validation accuracy and 4.35% loss. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Pulmonary Lung Cancer Classification Using Deep Neural Networks

Pulmonary Lung Cancer Classification Using Deep Neural Netwo...

引用

International Conference on Machine vision and Augmented Intelligence, MAI 2022

作者： Goswami, Jagriti Singh, Koushlendra Kumar Machine Vision and Intelligence Lab Department of Computer Science and Engineering National Institute of Technology Jamshedpur Jharkhand Jamshedpur831014 India

ISBN: (纸本)9789819901883

Lung cancer is the leading cause of cancer-related deaths globally. computer-assisted detection (CAD) systems have previously been used for various disease diagnosis and hence can serve as an efficient tool for lung cancer diagnosis. In this paper, we study the problem of lung cancer classification using chest computed tomography (CT) scans and positron emission tomography–computed tomography (PET-CT). A subset of publicly available Large-Scale CT and PET/CT Dataset for Lung Cancer Diagnosis (Lung-PET-CT-Dx) is used to train four different deep learning models using transfer learning for classifying three different types of lung cancer: Adenocarcinoma, Small Cell Carcinoma and Squamous Cell Carcinoma, by passing raw nodule patches to the network. The models are evaluated on metrics such as accuracy, precision, recall, F1-score and Cohen’s Kappa score. ROC curves and confusion matrices are also presented to provide a graphical representation of the models’ performance. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

Extending OpenStack Monasca for Predictive Elasticity Control

引用

Big Data Mining and Analytics 2024年第2期7卷 315-339页

作者： Giacomo Lanciano Filippo Galli Tommaso Cucinotta Davide Bacciu Andrea Passarella Scuola Normale Superiore Pisa 56126Italy the Real-Time Systems Laboratory(RETIS) TelecommunicationsComputer Engineeringand Photonics Institute(TeCIP)Scuola Superiore Sant’AnnaPisa 56127Italy Department of Computer Science University of PisaPisa 56127Italy National Research Council Pisa56127Italy

Traditional auto-scaling approaches are conceived as reactive automations,typically triggered when predefined thresholds are breached by resource consumption *** such rules at scale is cumbersome,especially when resources require non-negligible time to be *** paper introduces an architecture for predictive cloud operations,which enables orchestrators to apply time-series forecasting techniques to estimate the evolution of relevant metrics and take decisions based on the predicted state of the *** this way,they can anticipate load peaks and trigger appropriate scaling actions in advance,such that new resources are available when *** proposed architecture is implemented in OpenStack,extending the monitoring capabilities of Monasca by injecting short-term forecasts of standard *** use our architecture to implement predictive scaling policies leveraging on linear regression,autoregressive integrated moving average,feed-forward,and recurrent neural networks(RNN).Then,we evaluate their performance on a synthetic workload,comparing them to those of a traditional *** assess the ability of the different models to generalize to unseen patterns,we also evaluate them on traces from a real content delivery network(CDN)*** particular,the RNN model exhibites the best overall performance in terms of prediction error,observed client-side response latency,and forecasting *** implementation of our architecture is open-source.

关键词： OpenStack monitoring elasticity control auto-scaling predictive operations Monasca

来源：评论

学校读者我要写书评

暂无评论

A Low-Cost Pipelined Architecture Based on a Hybrid Sorting Algorithm

引用

IEEE Transactions on Circuits and Systems I: Regular Papers 2024年第2期71卷 717-730页

作者： Chen, You-Rong Ho, Chien-Chia Chen, Wei-Ting Chen, Pei-Yin The Digital Integrated Circuit Design Laboratory Department of Computer Science and Information Engineering National Cheng Kung University Tainan70101 Taiwan MediaTek Inc. Hsinchu30078 Taiwan

In this paper, a low-cost pipelined architecture based on a hybrid sorting algorithm is proposed. The proposed architecture is constructed with a bitonic sorter and several cascaded bidirectional insertion sorting units. The bidirectional insertion sorting unit uses the segmented sorted subsequence generated by the bitonic sorter as input, and records the maximum and minimum values of the subsequence. After all segmented subsequences are processed through the cascaded bidirectional insertion sorting units, a sorted sequence is obtained. The proposed architecture is implemented using the Verilog hardware description language (HDL) and synthesized using the Synopsys Design Compiler with a TSMC 90-nm cell library. The experimental results indicate that the proposed architecture can not only shorten sorting cycles but also reduce hardware area costs. Moreover, sorting cycles can be further shortened by increasing the parallelism of the proposed architecture. Under the configuration that 2048 32-bit data to be sorted and 16 data have to be processed simultaneously, the proposed architecture can improve the throughput-to-gate-count ratio by 16%, and throughput-to-power-consumption-ratio by 25% compared to the existing sorting design. The proposed architecture makes the most efficient use of hardware resources. © 2023 IEEE.

关键词： Costs

来源：评论

学校读者我要写书评

暂无评论

BiCSA-PUL: binary crow search algorithm for enhancing positive and unlabeled learning

引用

International Journal of Information Technology (Singapore) 2025年第3期17卷 1729-1743页

作者： Azizi, Nabil Ben Othmane, Mohamed Hamouma, Moumen Siam, Abderrahim Haouassi, Hichem Ledmi, Makhlouf Hamdi-Cherif, Aboubekeur Department of Computer Science Mostefa Ben Boulaid University Batna 2 Batna 05001 Algeria Department of Computer Science Abbes Laghrour University Khenchela 40004 Algeria Knowledge Engineering and IT Security Laboratory (ICOSI) Abbes Laghrour University Khenchela 40004 Algeria Department of Computer Science Ferhat Abbas University of Setif 1 Setif 19137 Algeria

This paper presents a novel metaheuristic binary crow search algorithm (CSA) designed for positive-unlabeled (PU) learning, a paradigm where only positive and unlabeled data are available, with applications in many diversified fields, such as medical diagnosis and fraud detection. The algorithm represent a useful adaptation of CSA, itself inspired by the food-hiding behavior of crows. The proposed BiCSA-PUL (binary crow search algorithm for positive-unlabeled learning) selects reliable negative samples from unlabeled data using binary vectors, and updates positions employing Hamming distance, guided by a modified F1-score, as fitness function. The algorithm was tested on 30 samples from 10 diverse datasets, outperforming seven state-of-the-art PU learning methods. The results reveal that BiCSA-PUL provides a robust and efficient approach for PU learning, significantly improving fitness and reliability. This work opens new avenues for applying metaheuristic optimization methods to challenging classification problems with limited labeled data. The main limitations are the potentially time-intensive process of parameters tuning and reliance on initial sampling. © Bharati Vidyapeeth's Institute of computer Applications and Management 2024.

关键词： Classification Machine learning Metaheuristic optimization Positive and unlabeled learning Raw search algorithm Two-step algorithms

来源：评论

学校读者我要写书评

暂无评论

Decentralizing Philippine Driver’s License Application and Identity Management for Transparency, Efficiency, and Security: A Design Proposal 4th

Decentralizing Philippine Driver’s License Application and...

引用

4th International Conference on Novel and Intelligent Digital Systems, NiDS 2024

作者： Lorico, Hans Daniel Q. Zamuco, Amiel David P. Caro, Jaime D. L. Juayong, Richelle Ann B. Service Science and Software Engineering Laboratory Department of Computer Science College of Engineering University of the Philippines Diliman Quezon City1101 Philippines

ISBN: (纸本)9783031733437

Blockchain is a developing and promising field in transaction and identity management. Recent efforts have been underway to address issues of data insecurity and inefficiency presented by centralized systems. Philippine government agencies, like the Land Transportation Office (LTO), traditionally store information in centralized databases, which can be susceptible to unauthorized access. This study makes the case of decentralizing certain LTO online procedures, starting with driver’s license application, as well as identity management using blockchain technology. With an immutable, distributed ledger system, it aims to introduce security, transparency, and efficiency into the system through a proof-of-concept web application. It demonstrates that not only certain LTO procedures can be decentralized, but also the management of license ID, more specifically the selective disclosure of information to requesting parties. The design presented here lays the groundwork for developing a decentralized application through Hyperledger Fabric and Indy, leveraging features like consensus mechanisms, verifiable credentials (VCs), and self-sovereign identity (SSI). The initial evaluation of the design using Hyperledger Caliper sets the stage for further research on decentralization, scalability, and efficiency. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Efficiency

来源：评论

学校读者我要写书评

暂无评论

Efficient and Lightweight Framework for Confidential Medical Image Search using Edge Computing 2

Efficient and Lightweight Framework for Confidential Medical...

引用

2nd International Conference on Computational Intelligence and Network Systems, CINS 2024

作者： Rajan, Arun Amaithi Vetriselvi, V. Aishwarya, R. Rajan, Anitha Amaithi Security Research Laboratory Department of Computer Science and Engineering College of Engineering Guindy Anna University Tamil Nadu Chennai600025 India Department of Computer Science and Business Systems Francis Xavier Engineering College Tamil Nadu Tirunelveli627011 India

ISBN: (纸本)9798331504106

In healthcare, medical images are playing a major role in accurate diagnosis. Cloud storage for medical images creates security risks and latency due to some critical vulnerabilities and their distance from data sources. By bringing computation closer to the source, edge computing improves efficiency, still edge servers cannot be trusted blindly, requiring robust security measures. In this article, an efficient and lightweight framework for secure medical image search and retrieval that leverages edge computing is proposed. Here, Medical images are encrypted and similarity-preserving hashcodes are generated locally at the source, which are sent to a trusted master edge server. Before being stored, searchable hashcodes undergo homomorphic encryption, allowing for secure computations on encrypted data. The master edge server generates n shares of the encrypted images by applying a Learning With Errors (LWE) based image sharing scheme and distributing them across nearby edge devices. During retrieval, to ensure resilience against device compromise, m out of n shares are used to reconstruct the image. The master edge server retrieves and reconstructs the top k relevant images in response to the query image. This framework enhances security and efficiency by reducing reliance on cloud storage and enabling secure processing on edge servers, making it ideal for modern healthcare environments. The efficiency and security of the proposed framework are proved theoretically and experimentally. Hence, it proved that the framework is performing better than existing state-of-the-art models. ©2024 IEEE.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

Learning to see speckle in the weak laser field through multimode fiber

引用

Optoelectronics Letters 2025年

作者： JI Yunqi SONG Binbin LI Xueqing LI Yonghui The Engineering Research Center of Learning-Based Intelligent System (Ministry of Education)， the Key Laboratory of Computer Vision and Systems (Ministry of Education)， and the School of Computer Science and Engineering， Tianjin University of Technology

Multimode fibers (MMFs) have great potential for endoscopic imaging due to the high number of modes and a small core diameter. Deep learning based on neural networks has received increasing attention in the field of scattering image reconstruction. However， most studies focus on designing complex network architectures to improve reconstruction， but these network models struggle to reconstruct images in a weak laser field. In the paper， a lightweight generative adversarial network model combined with a histogram specification algorithm is designed to reconstruct speckles in the weak laser field through MMF. Experimental results show that the reconstruction results of our algorithm have better metrics. Moreover， the model demonstrates excellent cross-domain generalization ability with regards to the Fashion-MNIST dataset. It is worth mentioning that we found that the speckles after inactivation still retain the ability to be reconstructed， which enhances the robustness of the model

关键词：

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：