检索结果-内蒙古大学图书馆

Mediterranean Conference on control and Automation (MED)

作者： R. Bustan R. Boriceanu D. Popescu L. Ichim Faculty of Automatic Control and Computer Science National University of Science and Technology POLITEHNICA Bucharest Bucharest Romania

ISBN: (数字)9798350395440

ISBN: (纸本)9798350395457

Face emotion recognition is an important subject in fields such as psychology and cognitive sciences, as well as in applications that use machine learning methods to create intelligent models capable of understanding human behavior. This paper addresses the challenge of negative facial emotion recognition by employing a comprehensive framework. Our approach leverages 68 facial landmarks and integrates three pre-trained convolutional neural network models (VGG19, DenseNet201, Inception V3) into a unified weighted voting system. The efficacy of our system is evaluated using the Extended Cohn-Kanade (CK+) and Karolinska Directed Emotional Faces (KDEF) databases. The results showcase the effectiveness of our approach in accurately identifying negative facial emotions. This research contributes to the advancement of emotion recognition systems., offering a promising methodology with implications for fields such as psychology., cognitive sciences., and the broader spectrum of intelligent systems reliant on facial expression analysis.

关键词： Human computer interaction Deep learning Emotion recognition Face recognition Neural networks Mental health Robustness

来源：评论

学校读者我要写书评

暂无评论

Tobacco Plant Detection Using a Performant Neural Network

Tobacco Plant Detection Using a Performant Neural Network

引用

International Conference on System Theory, control, and Computing (ICSTCC)

作者： Andrei Vicol Dan Popescu Loretta Ichim Faculty of Automatic Control and Computer Science National University of Science and Technology Politehnica Bucharest Bucharest Romania

ISBN: (数字)9798350364293

ISBN: (纸本)9798350364309

Accurate plant detection is an important task in modern precision agriculture, enabling growers to quantify their plant numbers and estimate crop yields accurately. This paper looks at an example tobacco crop with a small dataset and aims to augment it to produce reliable object detection and counting. To obtain a trainable dataset, the source masks are separated into base classes, noise is removed from the masks, and images are augmented. The newly obtained clean dataset is then split into patches and used to obtain a JSON file describing all the labels present in the image. The model used for this job is YOLOv8 and it is trained on a COCO-formatted augmented dataset. Choosing the proper set of weights is done by evaluating and studying the performance metrics over the whole training phase, picking the one that offers the greatest balance between performance, training duration, and risk of induced overfitting due to over-learning.

关键词： Training Measurement Precision agriculture Plants (biology) Noise Neural networks Object detection Reliability Crop yield Overfitting

来源：评论

学校读者我要写书评

暂无评论

Healing IoT Data for Indoor Air Quality Using Artificial Intelligence

Healing IoT Data for Indoor Air Quality Using Artificial Int...

引用

E-Health and Bioengineering Conference (EHB)

作者： Radu Nicolae Pietraru Adriana Olteanu Alexandru Ştefan Velicu Robert-Alexandru Craciun Faculty of Automatic Control and Computer Science National University of Science and Technology Politehnica Bucharest Bucharest Romania

ISBN: (数字)9798331532147

ISBN: (纸本)9798331532154

This paper reviews the problems encountered with data sets recorded by IoT monitoring networks and proposes five data healing methods. The five proposed methods include classic approaches but also methods based on artificial intelligence algorithms for predicting values. Using a case study based on indoor air quality research, the performances of the five methods of filling the gaps in the recorded data are explored both from the point of view of the ease of evaluation and visualization of the data and from the point of view of the statistical properties of the series of values.

关键词： Visualization Reviews Multiprocessor interconnection Data visualization Prediction algorithms Filling Artificial intelligence Monitoring Indoor air quality Biomedical engineering

来源：评论

学校读者我要写书评

暂无评论

Post-Processing of Thresholding or Deep Learning Methods for Enhanced Tissue Segmentation of Whole-Slide Histopathological Images

Post-Processing of Thresholding or Deep Learning Methods for...

引用

18th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2025

作者： Marczyk, Michal Wrobel, Agata Merta, Julia Polanska, Joanna Department of Data Science and Engineering Silesian University of Technology Gliwice44-100 Poland Yale Cancer Center Yale School of Medicine New HavenCT06511 United States Faculty of Automatic Control Electronics and Computer Science Silesian University of Technology Gliwice44-100 Poland

Digital pathology allows for the efficient storage and advanced computational analysis of stained histopathological slides of various tissues. Tissue segmentation is a crucial first step of digital pathology aimed at eliminating background, pen markings, and other artifacts, reducing image size, and increasing the efficiency of further analysis. In most cases, color thresholding or deep learning models are used, but their effectiveness is reduced due to complex artifacts and huge color variations between slides. We propose a post processing method to increase the tissue segmentation performance of any initial segmentation algorithm. Using a set of 197 manually annotated histopathological images of breast cancer patients and 63 images of endometrial cancer patients, we tested our method with 3 thresholding techniques and 3 deep learning-based algorithms by calculating the Dice index, Jaccard index, precision, and recall. In both datasets, applying post processing increased precision and recall for thresholding methods and mostly precision for deep learning models. Overall, applying post-processing gave better tissue segmentation performance than initial segmentation methods, significantly increasing Dice and Jaccard indices. Our results proved that thanks to post-processing, the tissue segmentation pipeline is more robust to noises and artifacts commonly present in histopathological images. © 2025 by SCITEPRESS– science and Technology Publications, Lda.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Multifunctional control system to enhance the operational capability of the unmanned multirotor platform dedicated to the diagnostics of building structures

引用

Bulletin of the Polish Academy of sciences: Technical sciences 2025年第3期73卷

作者： Czyba, Roman Domin, Jaroslaw Górski, Marcin Silesian University of Technology Faculty of Automatic Control Electronics and Computer Science Gliwice Poland Silesian University of Technology Faculty of Electrical Engineering Gliwice Poland Silesian University of Technology Faculty of Civil Engineering Gliwice Poland

Drones have become common devices that are increasingly used, from toy mini drones to very advanced autonomous flying platforms used in various industries. The safety of using engineering structures and the need to shorten the time of carrying out diagnostic activities, and thus reduce costs, force the use of innovative tools to carry out an automated and unmanned process of assessing the technical condition of existing engineering structures, especially large-size ones, including bridges strengthened with external FRP (fibre reinforced polymer) composite overlays. There are conventional methods for assessing the technical condition of existing building structures conducted by inspectors and experts, involving the use of typical inspection and diagnostic tools. During the research, it was found that it is possible to conduct an automated and unmanned process of assessing the technical condition of engineering structures using an unmanned flying platform equipped with a thermal imaging diagnostic device. The uniqueness of the presented solution is that the drone is equipped with a docking module that allows the drone to be temporarily attached to the structure being diagnosed for the duration of diagnostic activities. To ensure appropriate conditions for structural diagnostics, a unique system of automated diagnostic processes and a multi-stage structure of the control system in all phases of operation were developed. The article presents a solution for the control system based on control consistent with the reference model. In particular, the focus was on the difficult and dangerous issue of altitude control during the docking phase. Simulation tests conducted in the MATLAB/Simulink environment, as well as laboratory and field tests on a real drone proved the correctness of the proposed solutions, which facilitated the undisturbed operation of the diagnostic module. © 2025 The Author(s).

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Person Recognition and Authenticity Through Facial Analysis Combining LSTM and Siamese Networks

Person Recognition and Authenticity Through Facial Analysis ...

引用

International Conference on control, Automation, Robotics and Vision (ICARCV)

作者： Ionel Gabriel Zahia Loretta Ichim Dan Popescu Faculty of Automatic Control and Computer Science National University of Science and Technology POLITEHNICA Bucharest Bucharest Romania

ISBN: (数字)9798331518493

ISBN: (纸本)9798331518509

The automatic recognition of real people is particularly important in many fields of activity. The article proposes a complex system for recognizing people taking into account possible fraud attacks. A Long Short-Term Memory neural network was used to detect possible frauds in the video sequences, and to recognize the person in the case after validation as a real person, two Siamese neural networks implemented with VGG 16 and ResNet 50 were tested. The obtained results were compared for Euclidean distance, Cosine distance, and for two training methods: Contrastive Loss and Triple Loss. The best results were obtained for VGG 16, Cosine distance, and Triple Loss.

关键词： Training Convolution Neural networks Video sequences Euclidean distance Fraud Convolutional neural networks Long short term memory Robots Identification of persons

来源：评论

学校读者我要写书评

暂无评论

(POSTER) Visual Transformer Models on UAV-Captured Images for Person Detection

(POSTER) Visual Transformer Models on UAV-Captured Images fo...

引用

International Conference on Distributed Computing in Sensor Systems (DCOSS)

作者： Ionel - Gabriel Zahia Dan Popescu Loretta Ichim Faculty of Automatic Control and Computer Science National University of Science and Technology POLITEHNICA Bucharest Bucharest Romania

ISBN: (数字)9798350369441

ISBN: (纸本)9798350369458

One of the most important applications of UAVs is person detection for security or rescue tasks. The goal of the proposed paper is to develop, experiment, and compare the performance of two new neural networks based on the transformer architecture, Detection Transformer and Vision Transformer. Two datasets were used, an own one for testing and COCO for learning. The results are promising to take into account the difficulties of person detection at a distance.

关键词： Measurement computer vision Visualization Neural networks computer architecture Object detection Transformers

来源：评论

学校读者我要写书评

暂无评论

ChatGPT vs Sarcasm - How Well do State of the Art Language Models Handle Sarcasm? 9

ChatGPT vs Sarcasm - How Well do State of the Art Language M...

引用

9th International Conference on control, Decision and Information Technologies, CoDIT 2023

作者： Alexandru-Costin, Bǎroiu Matu Ştefan, Trǎuşan Politehnica University of Bucharest Faculty of Automatic Control and Computer Science Bucharest060042 Romania Research Institute for Artificial Intelligence 'Mihai Draganescu' of the Romanian Academy Bucharest050711 Romania

ISBN: (纸本)9798350311402

The work presented in this paper highlights the ability of state-of-the-art Large Language Models to generate sarcasm. In the recent months, ChatGPT has become increasingly popular and has started a new wave of interest in AI, especially on the non-scientific community. The model has displayed incredible feats, capable of producing human level results in a variety of tasks. To this end, sarcasm generation comes as a natural evaluation of the model's performance, as it was proven to be a highly difficult task, even for humans. As such, ChatGPT was tested in three experimental setups, no instructions, instructed to be sarcastic and instructed to impersonate a sarcastic character. The results generated by the model were benchmarked with a previous version, GPT2. Quantitative and qualitative analyses of the results were conducted, with ChatGPT achieving the highest BLEU score and generating the best results, outclassing its predecessor. © 2023 IEEE.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

DESIGNING A FUZZY controlLER FOR THE AUTOMATION OF AN IRRIGATION SYSTEM

UPB Scientific Bulletin, Series C: Electrical Engineering an...

引用

UPB Scientific Bulletin, Series C: Electrical Engineering and computer science 2023年第1期85卷 65-76页

作者： Bostan, Alexandra Dobrescu, Radu Hossu, Daniela Hossu, Andrei Faculty of Automatic Control and Computer Science University POLITEHNICA of Bucharest Romania Automatic Control and Industrial Informatics Department University POLITEHNICA of Bucharest Romania

The paper presents a solution for an irrigation controller based on the fuzzy-logic methodology. First, it describes the general problem of irrigation. Then, it discusses the physical control model. The precious irrigation is great significance for arid and semiarid area. In this paper, a fuzzy control algorithm was proposed to make an optimal irrigation strategy based on the actual measured soil humidity. In this paper, the fuzzy control algorithm was introduced in detail which included the setting of input and output, the selection of membership function and the setting of fuzzy rules. Following the discussion and the formal presentation of the fuzzy controller, the paper provide examples that will show the simplicity in designing and constructing such a system and other advantages of using fuzzy logic in the feedback control problem. © 2023, Politechnica University of Bucharest. All rights reserved.

关键词： Irrigation

来源：评论

学校读者我要写书评

暂无评论

Network Management devices in an SDN environment 24

Network Management devices in an SDN environment

引用

24th International Conference on control Systems and computer science, CSCS 2023

作者： Hamad, Diyar Jamal Yalda, Khirota Gorgees Tapus, Nicolae Okumus, Ibrahim Taner Politehnica University of Bucharest Faculty of Automatic Control and Computer Science Bucharest Romania Erbil Polytechnic University Soran Technical College Erbil Iraq Computer Engineering Kahramanmaras Sutcu Imam University Kahramanmaras Turkey

ISBN: (纸本)9798350313390

Network optimization and continued availability depend on a number of capabilities that are part of network management. The Maintenance, operating and also offering a safeguarded interaction network is very complicated. It calls for the network operators to grapple with low-level vendor particular arrangements to execute the high degree network policies which are complicated. A method for monitoring networks with OpenFlow controller is presented in this paper in two separate functions for the same networks. Bandwidth utilization, Meter values, charts and statistics are provided by the method to extend controller monitoring capabilities. The method architecture and implementation will be introduced in order to present the feature set. Additionally, softswitches are used as a switch and Mininet to evaluate a virtualized network. This analysis shows whether Meters or ports value is better for network management. © 2023 IEEE.

关键词： Network management

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：