Navegando por Assunto "Deep learning"

Agora exibindo 1 - 12 de 12

Acesso aberto (Open Access)
Beam tracking using deep learning applied to 6G MIMO
(Universidade Federal do Pará, 2024-12-16) OLIVEIRA, Ailton Pinto de; KLAUTAU JÚNIOR, Aldebaro Barreto da Rocha; http://lattes.cnpq.br/1596629769697284
This work explores the application of machine learning to enhance beam tracking in 6G MIMO Vehicle-to-Infrastructure (V2I) communications. Beam tracking, essential for sustaining reliable mmWave connections, remains challenging due to the high mobility of vehicular environments and the significant overhead associated with millimeter wave MIMO beamforming. While beam selection has been extensively studied, ML-based beam tracking is relatively underexplored, largely due to the scarcity of comprehensive datasets. To bridge this gap, this study introduces a novel public multimodal dataset, designed in accordance with 3GPP requirements, which combines wireless channel data with multimodal sensor information. This dataset supports the evaluation of advanced data fusion algorithms specifically tailored to V2I scenarios. Furthermore, a custom recurrent neural network (RNN) architecture is proposed as a robust solution for effective beam tracking, leveraging temporal and multimodal data to address the challenges of dynamic vehicular communications.
Acesso aberto (Open Access)
Beam-selection otimizado por aprendizado de máquina : uma abordagem multimodal
(Universidade Federal do Pará, 2023-12-30) FERREIRA, Jamelly Freitas; GOMES, Diego de Azevedo; http://lattes.cnpq.br/5116561408505726; KLAUTAU JÚNIOR, Aldebaro Barreto da Rocha; http://lattes.cnpq.br/1596629769697284
This dissertation aims to investigate the use of machine learning models using multimodal data as input to optimize the Beam-Selection process in millimeter-wave based networks. The use of Deep Learning has intensified in different areas, and it is possible to obtaing performance equal or superior to human performance, so its use is also promising in wireless communication scenarios. This work used data from different sources, which proved to be convenient since it is possible to adjust the model according to the quality/availability of this data. After executing the experiments and obtaining the results, it was observed that it is possible to obtain significant performance in different metrics even with simpler data such as image and coordinate.
Acesso aberto (Open Access)
Classificação de regiões de desmatamento via imagens do satélite landsat no nordeste do Pará
(Universidade Federal do Pará, 2023-12-18) CANAVIEIRA, Luena Ossana; COSTA, João Crisóstomo Weyl Albuquerque; http://lattes.cnpq.br/9622051867672434
Acesso aberto (Open Access)
Deep learning in education 5.0: proposing 3d geometric shapes classification model to improve learning on a metaverse application
(Universidade Federal do Pará, 2024-01-18) SANTOS, Adriano Madureira dos; SERUFFO, Marcos César da Rocha; http://lattes.cnpq.br/3794198610723464; https://orcid.org/0000-0002-8106-0560
The Brazilian educational system faces significant challenges, as evidenced by low educational development assessment scores. Due to the traditional educational model employed in the country, there are difficulties in the effective transmission of complex content, leading to high rates of academic failure and subsequent school dropout. The lack of innovation, especially in basic education settings, contributes to a scenario of low mathematical proficiency among Brazilian students. In this context, this work arises as a result of an innovation built to enhance the Geometa application, developed by the Inteceleri company, through the integration of Metaverse and Artificial Intelligence technologies to create an immersive and interactive educational environment. The intention is to train Artificial Intelligence for real-time three-dimensional geometric shape recognition from real-world object images. The proposal aims to mitigate challenges faced in Brazilian basic Mathematics education by adopting innovative technological approaches aligned with Education 5.0, which can be replicated for similar technologies involving the Metaverse. Furthermore, it is also intended to create a dynamic and sustainable educational environment that not only facilitates the mathematical concepts understanding but also promotes active student participation, encouraging their creativity and autonomy in the learning process. The method used relies on the ObjectNet dataset image reclassification from objects to three-dimensional geometric shapes. The reclassified images are used to train CNN, MobileNet, ResNet, ResNeXt, ViT and BEiT Deep Learning models, which are subsequently evalua ted through Machine Learning, inference time and dimension performance measures. Thus, the best-performance Artificial Intelligence model is selected for future integration into Geometa. As contributions of this work, the following were accomplished: (i) the defined models were trained for the three-dimensional geometric shapes recognition; (ii) the models were evaluated through Machine Learning, inference time and dimension performance measures; and (iii) the best-performance model was selected considering the highest assertiveness and smoothness based on models performances analysis. Concerning the obtained results, the ResNet surpassed BEiT, which was the second better performance model, in 5% Precision and 5 Inference Per Second. Finally, the ResNet model reached 84% Precision and 9 Inferences Per Second, being observed as the best-performance Artificial Intelligence for Geometa application integration flow.
Acesso aberto (Open Access)
Deep learning software-based holdover for PTP IEEE 1588 synchronization in 5G networks
(Universidade Federal do Pará, 2023-03-28) DUTRA, Rodrigo Gomes; KLAUTAU JÚNIOR, Aldebaro Barreto da Rocha; http://lattes.cnpq.br/1596629769697284
This work proposes evaluates software-based algorithm mechanisms for maintaining the synchronization of a real-time clock in holdover operation when the timing reference input is unavailable. Three algorithms, Autoregressive Integrated Moving Average (ARIMA), long short term memory (LSTM), and Transformer networks, are implemented and trained using timestamps and temperature data acquired while the slave clock is locked to a master clock. When the slave clock loses its reference, the algorithm-based models take over and control the clock. The proposed method is evaluated on a testbed of IEEE 1588 Precision Time Protocol (PTP) clocks based on field-programmable gate arrays, where nanosecond-accurate timestamps are collected for offline analysis. The models are evaluated using two clocks, one cost-effective, cristal oscillator (XO), and one robust, oven controlled cristal oscillator (OCXO), in both constant and variable temperature scenarios. The results show that all algorithms can sustain clock synchronization accuracy within reasonable Time division duplex (TDD) synchronization limits over intervals of 1000 seconds in all temperature and clock scenarios, with the transformerbased holdover mechanism outperforming the statistical approach and LSTM network. This cost-effective software-based approach proves to be feasible for increasing clock accuracy during holdover operation and can be generalized to other holdover contexts, such as in a Global Navigation Satellite System (GNSS) scenario.
Acesso aberto (Open Access)
Detecção de erosão em taludes baseada em deep learning
(Universidade Federal do Pará, 2023-03-31) LIMÃO, Caio Henrique Esquina; FRANCÊS, Carlos Renato Lisboa; http://lattes.cnpq.br/7458287841862567
The recent catastrophes triggered by the rupture of the Fundão and Córrego do Feijão dams caused around 300 deaths and countless irreparable socio-environmental damages. Since the use of more accurate monitoring systems and the proper execution of preventive and corrective maintenance would allow identifying, and even mitigating, the damage caused to society, it can be stated that there is a need for greater investment and incentive to create solutions of Structural Health Monitoring (SHM) capable of diagnosing occurrences that compromise the most crucial civil structures, such as bridges, buildings, dams and slopes. High-performance Artificial Intelligence (AI) techniques have been able to solve these structural analysis problems and presented superior results to previous solutions, their use has increased dramatically in the most diverse (SHM) scenarios. When it comes to image analysis and classification solutions, Convolutional Neural Network (CNN) is the type of neural network that delivers the best results. Therefore, this dissertation will describe the development process of a CNN with three convolutional layers that combines the use of the most consolidated technologies in the current scenario of computer vision, such as the Adam optimizer and batch normalization. The proposed CNN was trained with a database set up specifically for this dissertation, consisting of images of public work reports made by the Brazilian government, portfolios of companies that work with construction and maintenance of slopes and reports on landslides and/or catastrophes. These images were labeled, according to the context of each one of them, as stable or instable slopes. The results obtained were quite satisfactory, presenting an accuracy of 96.67% and proving that this solution is capable of identifying in a precise and improved way the instability indicators presented by the analyzed slopes, allowing a more adequate planning of the maintenance for each case, in the prevention of possible disasters, more efficient manpower management, cost reduction, greater safety and structural health to ensure its long-term integrity.
Acesso aberto (Open Access)
Detecção e rastreamento de componentes de vagões ferroviários utilizando redes neurais convolucionais e restricões geométricas
(Universidade Federal do Pará, 2020-04-27) GONÇALVES, Camilo Lélis Assis; BARROS, Fabrício José Brito; http://lattes.cnpq.br/9758585938727609
A inspeção de componentes de trem que podem causar descarrilamento possui um papel importante na manutenção ferroviária. A fim de aumentar a produtividade e a segurança, empresas prestadoras de serviços procuram por soluções de inspeção automáticas e confiáveis. Apesar da inspeção automática baseada em visão computacional ser um conceito consolidado, tais aplicações desafiam a comunidade de desenvolvimento em razão de fatores ambientais e logísticos a serem considerados. Este trabalho propõe uma técnica de detecção e estimativa das posições das regiões de dreno presentes em vagões de trem. Nosso detector/rastreador consiste em uma rede neural convolucional e um conjunto de restrições geométricas, que levam em conta a trajetória ideal dos componentes de interesse dos vagões e as distâncias entre eles. Detalhamos os procedimentos de treinamento e validação, juntamente com as métricas utilizadas para aferir a performance do sistema proposto. Os resultados apresentados são comparados com outras duas técnicas, e exibem um bom custo‑benefício entre confiança e complexidade computacional para a detecção dos componentes de interesse.
Acesso aberto (Open Access)
Estimação de descarga de dispositivo IoT usando deep learning com otimização NSGA-II
(Universidade Federal do Pará, 2024-02-28) MACEDO, Wilson Antonio Cosmo; BARROS, Fabrício José Brito; http://lattes.cnpq.br/9758585938727609
The increasing adoption of IoT (Internet of Things) network applications highlights the need to optimize energy management in these systems, because energy efficiency is crucial for the adaptability of IoT implementations. This study analyzes the discharge curves of a rechargeable battery in an IoT network context utilizing LoRa (Long Range) communication and various sensors, with the objective of generating multiple discharge curves to estimate the battery behavior in this scenario. These curves were used to train a Multilayer Artificial Neural Network (ANN), implementing Deep Learning techniques, where the ANN architecture was outlined using the NSGA-II (Non-dominated Sorting Genetic Algorithm II) Multi-objective Optimization algorithm. This resulted in models capable of estimating the battery discharge time by analyzing a segment of the discharge process observed by the model with a mean squared error of approximately two minutes for the most efficient model found. This result represents a very positive margin, considering that the duration of the discharge tests extends to approximately seventy-one hours and the data collection sampling rate is one minute.
Acesso aberto (Open Access)
Estimativa Volumétrica de Resíduos Sólidos Urbanos em Imagem de Visualização Única.
(Universidade Federal do Pará, 2024-09-02) AZANCORT NETO, Júlio Leite; FRANCÊS, Carlos Renato Lisboa; http://lattes.cnpq.br/7458287841862567
Efficient solid waste management is crucial for keeping the city clean and sustainable. This work presents a methodology that uses well-established algorithms for volume estimation in urban solid waste management from single-view images. The proposed system is based on state-of-the-art computer vision concepts and models, including instance segmentation, depth estimation, and volume calculation based on point clouds. The methodology demonstrated the ability to accurately estimate the volume of both individual and multiple solid waste objects in images. We evaluated our approach using real-world data. Despite challenges such as manual rescaling of distances and limited datasets, our system shows considerable potential for refinement and improvement, targeting complex scenarios like real urban environments. Numerical results indicated that the proposed system is promising even in complex scenarios, with mean absolute percentage errors (MAPE) of 8.60% for single waste and 9.23% for multiple wastes, resulting in an overall average of 8.91%. The coefficient of determination was 95.11% for single instances and 87.64% for multiple instances. The proposed methodology significantly contributes to the advancement of management technologies in smart cities.
Acesso aberto (Open Access)
Identificacao de larvas de mosquitos do genero aedes utilizando redes neurais convolucionais
(Universidade Federal do Pará, 2023-09-29) SILVA, Romário da Costa; FERREIRA JÚNIOR, José Jailton Henrique; http://lattes.cnpq.br/9031636126268760; FRANCÊS, Carlos Renato Lisboa
Arboviruses transmitted by mosquitoes of the Aedes genus constitute a threat to public health. Detection and control of these vectors are critical to preventing disease outbreaks including Dengue, Chikungunya, Zika and Yellow Fever. Computer vision and deep learning techniques have been increasingly used in epidemiological control, mainly with regard to the classification and detection of these mosquitoes. In this sense, three models are proposed for classification, detection and segmentation of mosquito larvae based on the use of convolutional neural networks (CNN) and object detection algorithms (YOLO). For this purpose, a dataset was created for training purposes. The dataset is composed of images of larvae, being categorized between Aedes and Non-Aedes classes. The results show that the proposed models are promising strategies and achieved accuracy values of 86.71%, mAP (Mean Average Precision) of 88.3% and 95.7% for the tasks of classification, detection and segmentation, respectively.
Acesso aberto (Open Access)
Machine learning algorithms for damage detection in structures under changing normal conditions
(Universidade Federal do Pará, 2017-01-31) SILVA, Moisés Felipe Mello da; SALES JÚNIOR, Claudomiro de Souza de; http://lattes.cnpq.br/4742268936279649; COSTA, João Crisóstomo Weyl Albuquerque; http://lattes.cnpq.br/9622051867672434
Engineering structures have played an important role into societies across the years. A suitable management of such structures requires automated structural health monitoring (SHM) approaches to derive the actual condition of the system. Unfortunately, normal variations in structure dynamics, caused by operational and environmental conditions, can mask the existence of damage. In SHM, data normalization is referred as the process of filtering normal effects to provide a proper evaluation of structural health condition. In this context, the approaches based on principal component analysis and clustering have been successfully employed to model the normal condition, even when severe effects of varying factors impose difficulties to the damage detection. However, these traditional approaches imposes serious limitations to deployment in real-world monitoring campaigns, mainly due to the constraints related to data distribution and model parameters, as well as data normalization problems. This work aims to apply deep neural networks and propose a novel agglomerative cluster-based approach for data normalization and damage detection in an effort to overcome the limitations imposed by traditional methods. Regarding deep networks, the employment of new training algorithms provide models with high generalization capabilities, able to learn, at same time, linear and nonlinear influences. On the other hand, the novel cluster-based approach does not require any input parameter, as well as none data distribution assumptions are made, allowing its enforcement on a wide range of applications. The superiority of the proposed approaches over state-of-the-art ones is attested on standard data sets from monitoring systems installed on two bridges: the Z-24 Bridge and the Tamar Bridge. Both techniques revealed to have better data normalization and classification performance than the alternative ones in terms of false-positive and false-negative indications of damage, suggesting their applicability for real-world structural health monitoring scenarios.
Acesso aberto (Open Access)
Metodologia para a classificação automática de doenças em plantas utilizando redes neurais convolucionais.
(Universidade Federal do Pará, 2019-11-07) REZENDE, Vanessa Castro; OLIVEIRA, Roberto Célio Limão de; http://lattes.cnpq.br/4497607460894318; SANTOS, Adam Dreyton Ferreira dos; http://lattes.cnpq.br/2616572481839756
Convolutional neural networks (CNNs) are one of the deep learning techniques that, due to the computational advancement of recent years, have enhanced the field of computer vision by enabling substantial gains in various classification problems, especially those involving digital images. Given the advantages of using these networks, a variety of applications for automatic plant diseases identification have been developed for specialized assistance or automated screening tools, contributing to more sustainable farming practices and improved food production security. In this context, this work aims to propose a methodology for the classification of multiple pathologies from distinct plant species, having as input a database composed of digital images of plant diseases. Initially, this methodology involved image preprocessing activities on the plant disease database to provide the appropriate input for selected CNN models (VGG16, RestNet101v1, ResNet101v2, ResNetXt50 and DenseNet169), as well as to generate ten new bases, ranging from 50 to 66 classes with greater representativeness, to submit the models to different situations. After model training, a comparative study was conducted based on widely used classification metrics such as test accuracy, f1-score, and area under the curve. To attest the significance of obtained results, the Friedman nonparametric statistical test and two post-hoc procedures were performed, which showed that ResNetXt50 and DenseNet169 obtained superior results when compared with VGG16 and ResNets.