检索结果-内蒙古大学图书馆

A fuzzy distance-based ensemb.e of deep models for cervical cancer detection

学校读者我要写书评

暂无评论

computer MeTHODS AND PROGRAMS IN bIOMeDICINe 2022年 219卷 106776-106776页

作者： Pramanik, Rishav biswas, Momojit Sen, Shibaprasad Junior, Luis Antonio de Souza Papa, Joao Paulo Sarkar, Ram [a]Department of Computer Science and Engineering Jadavpur University 188 Raja S C Mallick Rd Kolkata 700032 West Bengal India [b]Department of Metallurgical and Material Engineering Jadavpur University 188 Raja S C Mallick Rd Kolkata 700032 West Bengal India [c]Department of Computer Science and Technology University of Engineering and Management Kolkata 700160 West Bengal India [d]Department of Computing São Carlos Federal University-UFScar São Carlos São Paulo Brazil [e]Regensburg Medical Image Computing (ReMIC) Ostbayerische Technische Hochschule Regensburg (OTH Regensburg) Regensburg Bavaria Germany [f]Department of Computing São Paulo State University Av. Eng. Luiz Edmundo Carrijo Coube 14-01 Bauru São Paulo Brazil

background and Ob.ective: Cervical cancer is one of the leading causes of women's death. Like any other disease, cervical cancer's early detection and treatment with the best possib.e medical advice are the paramount steps that should be taken to ensure the minimization of after-effects of contracting this disease. PaP smear images are one the most effective ways to detect the presence of such type of cancer. This article proposes a fuzzy distance-based ensemb.e approach composed of deep learning models for cervical cancer detection in PaP smear ***: We employ three transfer learning models for this task: Inception V3, MobileNet V2, and Inception ResNet V2, with additional layers to learn data-specific features. To aggregate the outcomes of these models, we propose a novel ensemb.e method based on the minimization of error values between the ob.erved and the ground-truth. For samples with multiple predictions, we first take three distance measures, i.e., euclidean, Manhattan (City-block), and Cosine, for each class from their corresponding best possib.e solution. We then defuzzify these distance measures using the product rule to calculate the final ***: In the current experiments, we have achieved 95.30%, 93.92%, and 96.44% respectively when Inception V3, MobileNet V2, and Inception ResNet V2 run individually. After applying the proposed ensemb.e technique, the performance reaches 96.96% which is higher than the individual ***: experimental outcomes on three publicly availab.e datasets ensure that the proposed model presents competitive results compared to state-of-the-art methods. The proposed approach provides an end-to-end classification technique to detect cervical cancer from PaP smear images. This may help the medical professionals for better treatment of the cervical cancer. Thus increasing the overall efficiency in the whole testing process. The source code of the proposed work can be found in ***/ rishavprama

关键词： Cervical cancer computer-aided detection Deep learning Fuzzy logic ensemb.e learning

enabling dynamic and intelligent workflows for HPC, data analytics, and AI convergence

学校读者我要写书评

暂无评论

FUTURe GeNeRATION computer SYSTeMS-THe INTeRNATIONAL JOURNAL OF escience 2022年 134卷 414-429页

作者： ejarque, Jorge badia, Rosa M. Albertin, Loic Aloisio, Giovanni baglione, enrico becerra, Yolanda boschert, Stefan berlin, Julian R. D'Anca, Alessandro elia, Donatello exertier, Francois Fiore, Sandro Flich, Jose Folch, Arnau Gibbons, Steven J. Koldunov, Nikolay Lordan, Francesc Lorito, Stefano Lovholt, Finn Macias, Jorge Marozzo, Fabrizio Michelini, Alberto Monterrubio-Velasco, Marisol Pienkowska, Marta de la Puente, Josep Queralt, Anna Quintana-Orti, enrique S. Rodriguez, Juan e. Romano, Fabrizio Rossi, Riccardo Rybicki, Jedrzej Kupczyk, Miroslaw Selva, Jacopo Talia, Domenico Tonini, Roberto Trunfio, Paolo Volpe, Manuela [a]Barcelona Supercomputing Center (BSC) Spain [b]Centre Internacional de Mètodes Numèrics a l’Enginyeria (CIMNE) Spain [c]Universitat Politècnica de Catalunya (UPC) Spain [d]Jülich Supercomputing Centre (JSC) Germany [e]Universitat Politècnica de València (UPV) Spain [f]Atos BDS R&D HPC & AI Software France [g]DtoK Lab Srl Italy [h]Centro Euro-Mediterraneo sui Cambiamenti Climatici (CMCC) Italy [i]Department of Information Engineering and Computer Science University of Trento Italy [j]Universidad de Málaga (UMA) Spain [k]Istituto Nazionale di Geofisica e Vulcanologia (INGV) Italy [l]Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung Germany [m]Consejo Superior Investigaciones Cientificas (CSIC) Spain [n]Eidgenössische Technische Hochschule (ETH) Zürich Switzerland [o]Siemens AG Germany [p]Norwegian Geotechnical Institute (NGI) Norway [q]Poznan Supercomputing and Networking Center (PSNC) Poland

The evolution of High-Performance Computing (HPC) platforms enab.es the design and execution of progressively larger and more complex workflow applications in these systems. The complexity comes not only from the number of elements that compose the workflows but also from the type of computations they perform. While traditional HPC workflows target simulations and modelling of physical phenomena, current needs require in addition data analytics (DA) and artificial intelligence (AI) tasks. However, the development of these workflows is hampered by the lack of proper programming models and environments that support the integration of HPC, DA, and AI, as well as the lack of tools to easily deploy and execute the workflows in HPC systems. To progress in this direction, this paper presents use cases where complex workflows are required and investigates the main issues to be addressed for the HPC/DA/AI convergence. based on this study, the paper identifies the challenges of a new workflow platform to manage complex workflows. Finally, it proposes a development approach for such a workflow platform addressing these challenges in two directions: first, by defining a software stack that provides the functionalities to manage these complex workflows;and second, by proposing the HPC Workflow as a Service (HPCWaaS) paradigm, which leverages the software stack to facilitate the reusability of complex workflows in federated HPC infrastructures. Proposals presented in this work are sub.ect to study and development as part of the euroHPC eFlows4HPC project. (C) 2022 elsevier b.V. All rights reserved.

关键词： High performance computing Distributed computing Parallel programming HPC-DA-AI convergence Workflow development Workflow orchestration

A novel word sense disambiguation approach using wordnet knowledge graph

学校读者我要写书评

暂无评论

arXiv 2021年

作者： AlMousa, Mohannad benlamri, Rachid Khoury, Richard Department of Software Engineering Lakehead University Thunder BayONP7B 5E1 Canada Department of Computer Science and Software Engineering Université Laval QuébecQCG1V 0A6 Canada

Various applications in computational linguistics and artificial intelligence rely on high-performing word sense disambiguation techniques to solve challenging tasks such as information retrieval, machine translation, question answering, and document clustering. While text comprehension is intuitive for humans, machines face tremendous challenges in processing and interpreting a human’s natural language. This paper presents a novel knowledge-based word sense disambiguation algorithm, namely Sequential Contextual Similarity Matrix Multiplication (SCSMM). The SCSMM algorithm combines semantic similarity, heuristic knowledge, and document context to respectively exploit the merits of local context between consecutive terms, human knowledge about terms, and a document’s main topic in disambiguating terms. Unlike other algorithms, the SCSMM algorithm guarantees the capture of the maximum sentence context while maintaining the terms’ order within the sentence. The proposed algorithm outperformed all other algorithms when disambiguating nouns on the combined gold standard datasets, while demonstrating comparab.e results to current state-of-the-art word sense disambiguation systems when dealing with each dataset separately. Furthermore, the paper discusses the impact of granularity level, ambiguity rate, sentence size, and part of speech distribution on the performance of the proposed algorithm. © 2021, CC bY.

关键词： Semantics

Modeling user choice behavior under data corruption: Robust learning of the latent decision threshold model

学校读者我要写书评

暂无评论

A I I E Transactions

A I I e Transactions 2024年第12期56卷

作者： Feng Lin Xiaoning Qian bobak Mortazavi Zhangyang Wang Shuai Huang Cynthia Chen a Department of Industrial & Systems Engineering University of Washington Seattle WA USA b Department of Electrical & Computer Engineering Texas A&M University College Station TX USA c Department of Computer Science & Engineering Texas A&M University College Station TX USA d Department of Electrical & Computer Engineering University of Texas at Austin Austin TX USA e Department of Civil & Environmental Engineering University of Washington Seattle WA USA

Recent years have witnessed the emergence of many new mobile apps and user-centered systems that interact with users by offering choices with rewards. These applications have b.en promising to address challenging societal prob.ems such as congestion in transportation and behavior changes for healthier lifestyles. Considerab.e research efforts have b.en devoted to model the user behaviors in these new applications. However, as real-world user data is often prone to data corruptions, the success of these models hinges on a robust learning method. building on the recently proposed Latent Decision Threshold model, this article shows that, among the existing robust learning frameworks, theL0-norm-based framework can outperform other state-of-the-art methods in terms of prediction accuracy and model estimation. And based on theL0-norm framework, we further develop a user screening algorithm to identify potential bad actors.

关键词： Choice behavior modeling latent decision threshold model robust learning data corruption bad actor detection

Lesion Synthesis for Robust Segmentation of Infected Lung Region on Small-Scale Data

学校读者我要写书评

暂无评论

SSRN

SSRN 2022年

作者： Kadia, Dhaval Nguyen, Tam V. Asari, Vijayan University of Dayton 300 College Park DaytonOH United States 515 E Vista Del Cerro Dr TempeAZ85281 United States Jessie Hathcock 141 Department of Computer Science University of Dayton 300 College Park Ave DaytonOH45469 United States Kettering Laboratories Room 261B Department of Electrical and Computer Engineering University of Dayton 300 College Park Ave DaytonOH45469 United States

background and Ob.ective: Data-driven Artificial Intelligence (AI) has grown exponentially with novel computational architectures and cognitive capabilities to help solve complex prob.ems in medical imaging. Lung segmentation is essential since it processes the volumetric information of the lungs, removes the unnecessary areas of the scan. Deep learning eliminates feature engineering by progressively extracting complex high-level features from training data. both deep neural network (DNN) and the availability of diverse annotated data make the deep learning based solution robust and generalized for clinical use. even if having sophisticated DNN, scarcity of annotated data challenges the expected outcomes. Accurate segmentation of infected lungs with COVID-19 or any infection requires rich labeled data. Accurate pixel-level annotation tasks to generate such data are time-consuming and delay data ***: We propose a novel algorithm to generate synthetic lesion, or in other terms, infection-like artificial patterns, to help U-Net based DNN for robust infected lung and region segmentation and address data scarcity. Our lesion synthesis algorithm generates 2D and 3D patterns to create an enormous amount of synthetic data. We trained deep learning models on such synthetic data, tested them on actual data, and ob.erved promising ***: The proposed research helps the DNN generalize on a given domain to accomplish state-of-the-art segmentation results in the ab.ence of exact data. We developed and tested 2D and 3D deep learning models and achieved the average dice similarity coefficient of 0.9711 and 0.664 for infected lung segmentation and infected region segmentation, ***: There is much research on modern data-driven neural network based synthesis methods for infected lung region segmentation, whereas our significant contribution is proposing a classical synthesis algorithm to do the same. A key benefit of our algorithm is that

关键词： biological organs

Nonlinear ob.ervers design for vision-aided inertial navigation systems

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Wang, Miaomiao berkane, Soulaimane Tayebi, Ab.elhamid The Department of Electrical and Computer Engineering Western University LondonONN6A 3K7 Canada The Department of Department of Computer Science and Engineering University of Quebec in Outaouais QCJ8X 3X7 Canada The Department of Electrical Engineering Lakehead University Thunder BayONP7B 5E1 Canada

This paper deals with the simultaneous estimation of the attitude, position and linear velocity for vision-aided inertial navigation systems. We propose a nonlinear ob.erver on SO(3) × R15 relying on body-frame acceleration, angular velocity and (stereo or monocular) bearing measurements of some landmarks that are constant and known in the inertial frame. Unlike the existing local Kalman-type ob.ervers, our proposed nonlinear ob.erver guarantees almost global asymptotic stability and local exponential stability. A detailed uniform ob.ervability analysis has b.en conducted and sufficient conditions are derived. Moreover, a hybrid version of the proposed ob.erver is provided to handle the intermittent nature of the measurements in practical applications. Simulation and experimental results are provided to illustrate the effectiveness of the proposed state ob.erver. Copyright © 2021, The Authors. All rights reserved.

关键词： Inertial navigation systems

Resource allocation in STAR-RIS-Aided networks: OMA and NOMA

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Wu, Chenyu Mu, Xidong Liu, Yuanwei Gu, Xuemai Wang, Xianbin Harbin150001 China School of Artificial Intelligence Beijing University of Posts and Telecommunications Beijing100876 China School of Electronic Engineering and Computer Science Queen Mary University of London LondonE1 4NS United Kingdom Department of Electrical and Computer Engineering Western University LondonONN6A 5B9 Canada

Simultaneously transmitting and reflecting reconfigurab.e intelligent surface (STAR-RIS) is a promising technology to achieve full-space coverage by splitting the incident signal into the transmitted and reflected signals towards both sides of the surface. This paper investigates the resource allocation prob.em in a STAR-RIS-assisted multi-carrier communication network. To maximize the system sum-rate, a joint optimization prob.em of channel assignment, power allocation, and transmission- and reflection-beamforming at the STAR-RIS for orthogonal multiple access (OMA) is first formulated, which is a mixed-integer non-linear programming prob.em. To solve this challenging prob.em, we first propose a channel assignment scheme utilizing matching theory and then invoke the alternating optimization-based method to optimize the resource allocation policy and beamforming vectors iteratively. Furthermore, the sum-rate maximization prob.em for non-orthogonal multiple access (NOMA) with flexib.e decoding orders is investigated. To efficiently solve it, we first propose a location-based matching algorithm to determine the sub-channel assignment, where a transmitted user and a reflected user are grouped on a sub-channel. based on this transmission-and-reflection sub-channel assignment strategy, a three-step approach is proposed, where the decoding orders, beamforming-coefficient vectors, and power allocation are optimized by employing semidefinite programming, convex upper bound approximation, and geometry programming, respectively. Numerical results unveil that: 1) For OMA, a general design that includes same-side user-pairing for channel assignment is preferab.e, while for NOMA, the proposed transmission-and-reflection scheme can achieve comparab.e performance as the exhaustive search-based algorithm. 2) The STAR-RIS-aided NOMA network significantly outperforms the networks employing conventional RISs and OMA. Copyright © 2021, The Authors. All rights reserved.

关键词： Iterative methods

Matching evaluation based on image content discriminative features for different image types

学校读者我要写书评

暂无评论

The Journal of Photographic science 2024年第1期72卷

作者： eman S. Sabry Salah elagooz Fathi e. Abd el-Samie Nirmeen A. el-bahnasawy Ghada M. el-banby Rab.e A. Ramadan a Department of Communications and Computers Engineering Higher Institute of Engineering El-Shorouk Academy El-Shorouk City Egypt b Department Electronics and Electrical Communications Engineering Faculty of Electronic Engineering Menoufia University Menouf Egyptc Department of Information Technology College of Computer and Information Sciences Princess Nourah Bint Abdulrahman University Riyadh Saudi Arabia d Computer Science and Engineering Department Faculty of Electronic Engineering Menoufia University Menouf Egypt e Department of Industrial Electronics and Control Engineering Faculty of Electronic Engineering Menoufia University Menouf Egypt f Computer Engineering Department College of Engineering Cairo University Giza Egypt

AbSTRACTSearch by image has taken the place of the prior wordy approach, considering spatial image comparison in the feature space. Several query image forms are availab.e. Sketch-to-real image matching is very challenging prob.em, which is commonly based on edge feature detectors to extract features from sketches. However, they are inconvenient, and it is crucial to select different methods. Thus, methods assessment is conducted for similarity matching between sketched and real images for different drawing contents and clarity levels. The assessment approach includes comparing feature keypoints gathered using the Scale-Invariant Feature Transform (SIFT) and Oriented FAST and Rotated bRIeF (ORb) methods. In addition, the comparison includes matching of ob.ects inside High-Deﬁnition (HD) images and 3D images based on their generated features. The textual-Hcontent image search is another search type that is frequently used in brand logo cases. Hence, it is presented for text-logo image search based on visual features. besides, the motives and downsides of the considered methods are declared, including the space and time complexities within test cases.

关键词： SIFT ORb sketch-real image retrieval ob.ects matching spatial distance 3D image matching

P-WAe: Generalized patch-wasserstein autoencoder for anomaly screening

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zhang, Hui Chen, Yurong Wang, Yaonan Cao, Yihong Wu, Q.M. Jonathan Yang, Yimin National Engineering Laboratory of Robot Visual Perception and Control Technology School of Robotics Hunan University Changsha Hunan410082 China Department of Electrical and Computer Engineering University of Windsor WindsorONN9B3P4 Canada Department of Computer Science Lakehead University OntarioP7B 5E1 Canada Vector Institute TorontoM5G 1M1 Canada

To mitigate the inspector’s workload and improve the quality of the product, computer vision-based anomaly detection (AD) techniques are gradually deployed in real-world industrial scenarios. Recent anomaly analysis benchmarks progress to generative models. The aim is to model the defect-free distribution so that anomalies can be classified as out-of-distribution samples. Nevertheless, there are two disturbing factors that need researchers and deployers to prioritize: (i) the simplistic prior latent distribution inducing limited expressive capability;(ii) the collapsed mutual-dependent features resulting in poor generalization. In this paper, we propose a novel Patch-wise Wasserstein Autoencoder (P-WAe) architecture to alleviate those challenges. In particular, a patch-wise variational inference model coupled with solving the jigsaw puzzle is designed, which is a simple yet effective way to increase the expressiveness and complexity of the latent manifold. This alleviates the blurry reconstruction prob.em. In addition, the Hilbert-Schmidt Independence Criterion (HSIC) bottleneck is introduced to constrain the over-regularization representation. Comprehensive experiments, conducted on the MVTec AD dataset, demonstrate the superior performance of our proposed method. © 2021, CC bY.

关键词： Anomaly detection