检索结果-内蒙古大学图书馆

A convolutional neural network framework with ConvNeXt and Grad-CAM for classification of oral potentially malignant disorders and oral squamous cell carcinoma

引用

Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology 2025年第1期139卷 e28-e29页

作者： C Saldivia-Siracusa AV Barros-da-Silva E Santos-Carlos-de-Souza C Mariano Pedros AL Damaceno-Araujo M Ajudarte-Lopes PA Vargas MG Quiles AC Ponce-de-Leon-Ferreira-de-Carvalho LP Kowalski AR Santos-Silva Departamento de Diagnóstico Oral Faculdade de Odontologia de Piracicaba Universidade Estadual de Campinas Piracicaba São Paulo Brazil Institute of Science and Technology Federal University of São Paulo São José dos Campos São Paulo Brazil Institute of Mathematics and Computer Sciences University of São Paulo São Carlos São Paulo Brazil Head and Neck Surgery Department University São Paulo Medical School São Paulo São Paulo Brazil

Background Early detection of oral squamous cell carcinoma (OSCC) is crucial for reducing its high mortality rates. Deep learning (DL) artificial models have shown promise in recognizing patterns in complex samples, which can be valuable for improving the diagnosis of malignant and potentially malignant oral lesions. Objective To use the convolutional neural network (CNN) deep learning algorithms to develop an automated classification model for OSCC and oral potentially malignant disorders (OPMD) recognition using clinical images, and to assess interpretability of the model's regions of interest through Grad-CAM visual representation. Methods A cross-sectional study was performed retrospectively including a dataset of 837 clinical images obtained from 3 Brazilian oral medicine services, consisting of 351 images of OPMD and 486 images of OSCC. The dataset was divided into training, validation, and test subsets with an 8:1:1 proportion. Two transfer learning strategies were applied, using frozen and fine-tuned weights of ImageNet ILSVRC and ISIC 2019 to initially pre-train the ConvNeXt convolutional neural network (CNN). Results Pretraining with ILSVRC yielded the best performance for the ConvNeXt CNN model, achieving a mean accuracy, precision, recall, and F1 score of 1.00 during training and 0.78 during testing. Further pretraining with ISIC 2019 resulted in slightly lower training scores (0.99), yet the test scores were comparable to those of the latter strategy. Conclusions The ConvNeXt CNN model demonstrated high performance in distinguishing OPMD and OSCC. Interpretability of obtained results was facilitated by implementing Grad-CAM assisted representation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Analysis of the brazilian research agencies using a multicriteria decision aid known as TODIM 13

Analysis of the brazilian research agencies using a multicri...

引用

13th Annual IEEE International Systems Conference, SysCon 2019

作者： Magalhaes, Rachel F. Rangel, Luis A. D. Sbruzzi, Elton F. Nascimento, Cairo L. Leles, Michel C. R. School of Industrial Metallurgical Engineering Universidade Federal Fluminense Volta Redonda-RJ27255-125 Brazil Division of Computer Science Instituto Tecnológico de Aeronáutica São José dos Campos-SP12228-900 Brazil Division of Electronic Engineering Instituto Tecnológico de Aeronáutica São José dos Campos-SP12228-900 Brazil Department of Technology Universidade Federal de São Josédel-Rei Ouro Branco-MG36420-000 Brazil

ISBN: (纸本)9781538683965

The objective of this work is to analyze the productivity of different Brazilian research promotion agencies by using the Method of Interactive Decision Making and Multicriteria (Acronym TODIM in Portuguese), which is based on the Prospect Theory. The proposed model classify of 22 Brazilian research agencies (two federal and twenty state agencies) under six different criteria. The result shows that some small state agencies are more efficient than bigger ones in terms of the funding/scholarship ratio. This suggests that the winner model could be further investigated to be promoted among the other agencies of this country. © 2019 IEEE.

关键词： Decision support systems

来源：评论

学校读者我要写书评

暂无评论

Impact angle control of local intense dB/dt variations during shock-induced substorms

arXiv

引用

arXiv 2021年

作者： Oliveira, Denny M. Weygand, James M. Zesta, Eftyhia Ngwira, Chigomezyo M. Hartinger, Michael D. Xu, Zhonghua Giles, Barbara L. Gershman, Dan J. Silveira, Marcos V.D. Souza, Vítor M. Goddard Planetary Heliophysics Institute University of Maryland Baltimore County BaltimoreMD United States Heliophysics Science Division NASA Goddard Space Flight Center GreenbeltMD United States Department of Earth Planetary and Space Sciences University of California Los Angeles Los AngelesCA United States Atmospheric and Space Technology Research Associates LLC LouisvilleCO United States Space Science Institute BoulderCO United States Center for Space Science and Engineering Research Virginia Tech BlacksburgVA United States Bradley Department of Electrical and Computer Engineering Virginia Tech BlacksburgVA United States National Institute of Aerospace HamptonVA United States National Institute for Space Research - INPE SP São José dos Campos Brazil

The impact of interplanetary shocks on the magnetosphere can trigger magnetic substorms that intensify auroral electrojet currents. These currents enhance ground magnetic field perturbations (dB/dt), which in turn generate geomagnetically induced currents (GICs) that can be detrimental to power transmission infrastructure. We perform a comparative study of dB/dt variations in response to two similarly strong shocks, but with one being nearly frontal, and the other, highly inclined. Multi-instrument analyses by the Time History of Events and Macroscale Interactions during Substorms (THEMIS) and Los Alamos National Laboratory spacecraft show that nightside substorm-time energetic particle injections are more intense and occur faster in the case of the nearly head-on impact. The same trend is observed in dB/dt variations recorded by THEMIS ground magnetometers. THEMIS all-sky imager data show a fast and clear poleward auroral expansion in the first case, which does not clearly occur in the second case. Strong field-aligned currents computed with the spherical elementary current system (SECS) technique occur in both cases, but the current variations resulting from the inclined shock impact are weaker and slower compared to the nearly frontal case. SECS analyses also reveal that geographic areas with dB/dt surpassing the thresholds 1.5 and 5 nT/s, usually linked to high-risk GICs, are larger and occur earlier due to the symmetric compression caused by the nearly head-on impact. These results, with profound space weather implications, suggest that shock impact angles affect the geospace driving conditions and the location and intensity of the subsequent dB/dt variations during substorm activity. © 2021, CC BY.

关键词： Magnetosphere

来源：评论

学校读者我要写书评

暂无评论

Neural Network to Control a Multiple Angles Kicker in RoboCup Small Size League

Neural Network to Control a Multiple Angles Kicker in RoboCu...

引用

IEEE Latin American Robotics Symposium, LARS

作者： Francisco A. B. Azevedo Daniela Vacarini Emilia Villani Marcos R. O. A. Maximo Aeronautical Engeneering in Aeronautics Institute of Technology (ITA) São José dos Campos SP Brazil National Higher French Institute of Aeronautics and Space (ISAE-Supaero) Mechanical Engineering in Aeronautics Institute of Technology Autonomous Computational Systems Lab (LAB-SCA) Computer Science Division Aeronautics Institute of Technology Mechatronics Engineering and aerospace systems Aeronautics Institute of Technology

ISBN: (数字)9780738111537

ISBN: (纸本)9781665419857

The Small Size League (SSL) is a robot soccer league of the Robot World Cup (RoboCup). A recent trend in hardware design in SSL involves multiple angles kicking mechanisms. Combining a multiple angles kicker with a dribbler, which consists of a spinning bar to manipulate the ball through friction, a robot can execute angled and curved shoots and passes. However, due to the complex physical interactions involved, determining how to control the kicking and dribbling devices in order to obtain the desired ball trajectory is not simple. The ball motion may be mathematically modelled by a nonlinear ordinary differential equation (ODE). The parameters needed to perform a desired curved kick are the solution of the inverse problem of the nonlinear ODE. Since this is hard to compute in real-time, this work proposes a neural network to solve this inverse problem. The network is trained using many simulated ball trajectories. Finally, we show simulation results to validate the proposed method.

关键词： Robots Trajectory Mathematical model Friction Mobile robots Biological neural networks Real-time systems

来源：评论

学校读者我要写书评

暂无评论

Biodegradability of Partially Reduced Nanographene Oxide by Human, Plant and Microbial Enzymes: Impact of Magnetic Nanoparticles

SSRN

引用

SSRN 2024年

作者： Kurapati, Rajendra Swetha, K. Sudeshna, Samantaray Silva, Filipa A.L.S. Silva, Filipa C. Freitas, Bruno Incorvia, Jean Anne C. Banerjee, Sourav Afonso Fernandes, José Ramiro Jayaraj, Arya Singh, N. Sadananda Magalhães, Fernão D. Pinto, Artur M. School of Chemistry Indian Institute of Science Education and Research Thiruvananthapuram Maruthimala PO Thiruvananthapuram India LEPABE - Laboratory for Process Engineering Environment Biotechnology and Energy Faculdade de Engenharia Universidade do Porto Rua Dr. Roberto Frias s/n Porto4200-180 Portugal ALiCE - Associate Laboratory in Chemical Engineering Faculdade de Engenharia Universidade do Porto Rua Dr. Roberto Frias s/n Porto4200-180 Portugal i3S - Instituto de Investigação e Inovação em Saúde Universidade do Porto Rua Alfredo Allen 208 Porto4200-180 Portugal INEB - Instituto de Engenharia Biomédica Universidade do Porto Rua Alfredo Allen 208 Porto4200-180 Portugal IPATIMUP - Instituto de Patologia e Imunologia Molecular da Universidade do Porto Universidade do Porto Porto4200-180 Portugal Department of Electrical and Computer Engineering University of Texas at Austin. AustinTX78712 United States CQVR – Centro de Química Vila Real Universidade de Trás-os-Montes e Alto Douro Portugal Physical Department University of Trás-os-Montes and Alto Douro Quinta dos Prados Vila Real5000-801 Portugal School of Biology Indian Institute of Science Education and Research Thiruvananthapuram Maruthimala PO Thiruvananthapuram India

Although graphene-based materials continuously expand their range of industrial and biomedical applications, understanding their long-term fate in the organisms and environment is still in its infancy. Herein we examine the biodegradation of partially reduced graphene oxide (p-rGOn) and magnetic nanoparticles (IONPs) embedded rGOn (@rGOn) by incubating with the human myeloperoxidase (hMPO) isolated from the neutrophils, including with plant enzyme, horseradish peroxidase and treating with microbial laccase in the presence of its redox mediator ABTS [2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)]. Further, the biodegradation of p-rGOn and @rGOn was investigated by incubating with the laccase-secreting fungi, Aspergillus sp, present in the compost soil. The biodegradation of both rGO samples was analyzed for morphological and structural changes using electron microscopy and Raman spectroscopy. The results confirmed the efficient degradation of p-rGOn and that the presence of IONPs enhanced the biodegradability of @rGOn (magnetic) over p-rGOn, which could be attributed to the Fenton-type reaction because of IONPs. These results confirmed that the presence of IONPs facilitates the magnetic manipulation of @rGOn and increases their biodegradability. The current work could be crucial for better designing the potential biomedical applications of @rGOn (magnetic) for near-infrared photothermal (PTT) and externally controlled magnetotherapy applications useful in cancer treatment. © 2024, The Authors. All rights reserved.

关键词： Enzymes

来源：评论

学校读者我要写书评

暂无评论

Rpas in the support for photogrammetry education: Cases in topographic mapping and documentation of historical monuments 4

Rpas in the support for photogrammetry education: Cases in t...

引用

4th ISPRS Geospatial Week 2019

作者： Reiss, M.L.L. Mendes, T.S.G. Andrade, M.R.M. Amory, A.M. De Lara, R. Souza, S.F. LAFOTO-Laboratory of Photogrammetry Research Department of Geodesy Institute of Geoscience UFRGS-Federal University of Rio Grande Do sul Brazil LAGEO-Laboratory of Geodesy Department of Geodesy Institute of Geoscience UFRGS-Federal University of Rio Grande Do sul Brazil Faculty of Computer Science PUCRS-Pontifical Catholic University of Rio Grande Do sul Brazil Department of Environmental Engineering Institute of Science and Technology-ICT University of São Paulo - UNESP São José Dos Campos São Paulo Brazil National Center for Monitoring and Early Warning of Natural Disasters-CEMADEN São José Dos Campos São Paulo Brazil

Currently on the world market there are several models and manufacturers of RPAS-UAV. This equipment made it possible to complete the entire flow of the photogrammetric mapping process in teaching for the students of the Cartographic Engineering course. In the Institute of Geosciences, this course already counted on the photogrammetric stations, equipped with hardware for 3D visualization (graphic card, high-speed monitors and stereoscopic vision glasses), photogrammetric software for data processing, and equipment for geodetic measurement control points. Before the acquisition of the RPAS, for the accomplishment of the academic works, it was necessary to buy or to look for in donations from companies of aerial photographs, that did not always correspond to an area of scientific interest. In addition, the scenes in the photos were not current, and it was impossible to have control points that were pre-signalized. From the availability of the RPAS in the University, the students began to have more interest by Photogrammetry in their professional end of coursework and throughout the disciplines. It became possible to carry out in practice all the processes involved in the mapping of nearby areas of interest, from the planning to the generation of photogrammetric products, such as the generation of DTM (Digital Terrain Model), DEM (Digital Elevation Model), orthophotomaps and extraction of features. In addition to products related to the 3D mapping of terrains, mainly through the RPAS of rotating wings, there was also the possibility of mapping and reconstructing 3D structures, such as buildings of historical monuments. © Authors 2019.

关键词： Photogrammetry

来源：评论

学校读者我要写书评

暂无评论

Learning Humanoid Robot Running Skills through Proximal Policy Optimization

arXiv

引用

arXiv 2019年

作者： Melo, Luckeciano C. Máximo, Marcos R.O.A. Computer Science Division Aeronautics Institute of Technology Praça Marechal Eduardo Gomes 50 Vila das Acácias São José dos CamposSP12228-900 Brazil

In the current level of evolution of Soccer 3D, motion control is a key factor in team’s performance. Recent works takes advantages of model-free approaches based on Machine Learning to exploit robot dynamics in order to obtain faster locomotion skills, achieving running policies and, therefore, opening a new research direction in the Soccer 3D environment. In this work, we present a methodology based on Deep Reinforcement Learning that learns running skills without any prior knowledge, using a neural network whose inputs are related to robot’s dynamics. Our results outperformed the previous state-of-the-art sprint velocity reported in Soccer 3D literature by a significant margin. It also demonstrated improvement in sample efficiency, being able to learn how to run in just few hours. We reported our results analyzing the training procedure and also evaluating the policies in terms of speed, reliability and human similarity. Finally, we presented key factors that lead us to improve previous results and shared some ideas for future work. Copyright © 2019, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

High Speed Operational Amplifier using a-InGaZnO TFTs with Negative Capacitance

High Speed Operational Amplifier using a-InGaZnO TFTs with N...

引用

IEEE International Symposium on Circuits and Systems (ISCAS)

作者： Ricardo Rodrigues Pydi Bahubalindruni Pedro Barquinha Department of Electrical Engineering and Computer Sciences Indian Institute of Science Education and Research Bhopal Bhopal Bypass Road Bhauri Bhopal 462 066 CENIMAT/I3N Departamento de Ciěncia dos Materiais Faculdade de Ciěncias e Tecnologia FCT Universidade Nova de Lisboa and CEMOP-UNINOVA 2829-516 Caparica Portugal

ISBN: (数字)9781728133201

ISBN: (纸本)9781728133218

This paper presents a novel high speed operational amplifier using a-IGZO TFTs for a given technology, without changing device structure, processing conditions and materials. In the proposed design, a negative capacitance generator (NCG) is employed, using a-IGZO TFTs, which is connected to the output of a positive feedback operational amplifier. This technique helps to move the output pole of the amplifier to very high frequency by reducing the overall equivalent capacitance (C eq ) at the output node, as the negative capacitance is in parallel with C eq . By using this technique, the unity gain bandwidth (GB) of the opamp is increased from 486 kHz to 1.474 MHz without compromising other performance metrics, such as, gain and stability (phase margin). However power consumption is increased from 0.3mW to 0.6 mW, when the simulations took place with in-house IGZO TFT models. Both NCG and the amplifier were designed with a minimum feature size of 10 μm and a power supply of 10V and 15V, respectively.

关键词： Capacitance Bandwidth Thin film transistors Gain Generators Power demand

来源：评论

学校读者我要写书评

暂无评论

People still care about facts: Twitter users engage more with factual discourse than misinformation-a comparison between COVID and general narratives on twitter

arXiv

引用

arXiv 2020年

作者： Silva, Mirela Ceschin, Fabrício Shrestha, Prakash Brant, Christopher Gilda, Shlok Fernandes, Juliana Silva, Catia S. Grégio, André Oliveira, Daniela Giovanini, Luiz Electrical & Computer Engineering University of Florida P.O. Box 116200 GainesvilleFL32611 United States Federal University of Paraná Cel. Francisco Heráclito dos Santos 100 Paraná Curitiba81530-000 Brazil Computer and Information Science & Engineering University of Florida P.O. Box 116200 GainesvilleFL32611 United States College of Journalism and Communications University of Florida P.O. Box 118400 GainesvilleFL32611 United States

Misinformation entails the dissemination of falsehoods that leads to the slow fracturing of society via decreased trust in democratic processes, institutions, and science. The public has grown aware of the role of social media as a superspreader of untrustworthy information, where even pandemics have not been immune. In this paper, we focus on COVID-19 misinformation and examine a subset of 2.1M tweets to understand misinformation as a function of engagement, tweet content (COVID-19-vs. non-COVID-19-related), and veracity (misleading or factual). Using correlation analysis, we show the most relevant feature subsets among over 126 features that most heavily correlate with misinformation or facts. We found that (i) factual tweets, regardless of whether COVID-related, were more engaging than misinformation tweets;and (ii) features that most heavily correlated with engagement varied depending on the veracity and content of the tweet. Copyright © 2020, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

On the UML use in the brazilian industry: A state of the practice survey 30

On the UML use in the brazilian industry: A state of the pra...

引用

30th International Conference on Software Engineering and Knowledge Engineering, SEKE 2018

作者： Farias, Kleinner Gonçales, Lucian Bischoff, Vinicius Da Silval, Bruno Guimarães, Everton Nogle, Jacob Univ. of Vale Do Rio Dos Sinos São Leopoldo Brazil Dept. of Computer Science and Software Engineering California Polytechnic State University San Luis Obispo United States Department of Computer Science Drexel University Philadelphia United States

ISBN: (纸本)1891706446

Context: The Unified Modeling Language (UML) has become the standard for modeling software. Several surveys on the UML usage have been proposed in recent years. However, none of them explores the UML use in specific regional scope, and thus little is known about the practices and perceptions of UML use from the perspective of practitioners in the Brazilian industry. Objective: This paper reports on a survey focused on identifying the state-of-The-practice of the Brazilian industry for what concerns the UML usage in real-world settings. Method: In total, 222 practitioners from 140 different Information Technology companies have answered an on-line (or printed) questionnaire concerning their UML use experiences, the difficulty in adopting UML and what should be done to increase the UML adoption in practice. Result: The results show that: (1) 60 participants (28.2%) have used UML in their daily work, while 73.2% have not;(2) 55.41% of the surveyed participants did not disagree with the statement that UML is the "lingua franca" in software modeling;(3) 61.26% reported to find that the automatic creation of UML diagrams to represent a big picture of the system under development would be useful to boost UML use. Conclusion: The UML is not often used in the work life of participants. In addition, no relationship was identified between the use of UML and the participant company being a software factory. © 2018 Universitat zu Koln. All rights reserved.

关键词： Unified Modeling Language

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：