检索结果-内蒙古大学图书馆

A novel hybrid deep learning framework for pedestrian detection based on thermal infrared and visible spectrum images

引用

Multimedia Tools and Applications 2025年 1-27页

作者： Defaoui, Mahassine Koutti, Lahcen El Ansari, Mohamed Lahmyed, Redouan Masmoudi, Lhoussaine LabSIV Department of Computer Science Faculty of Science Ibn Zohr University Agadir80000 Morocco Informatics and Applications Laboratory Faculty of Science My Ismail University Meknes Morocco Conception and Systems Laboratory Faculty of Sciences Mohamed V University Rabat Morocco

For many intelligent systems, designing accurate pedestrian detection approaches is a fundamental task. This paper describes a novel hybrid system to detect pedestrians using both visible and thermal infrared sensors. The proposed method is achieved in two primary stages. The first stage, hypotheses generation (HG), aims to extract the regions that represent suspected pedestrians from the image. It is performed using both thermal and visible images of the same scene. Here, an improved deep learning-based solution with a convolutional neural network named PedTherm-U-net is proposed. It segments the thermal images and maps the segmentation results into the corresponding visible image to extract the ROIs (candidate pedestrians). The second stage, hypotheses verification (HV), classifies the ROIs extracted from the first stage into pedestrians and non-pedestrians. In the HV step, a novel PedVis-ResNet framework to classify the extracted ROIs is introduced. It is a modified version of ResNet model dedicated to classify the visible pedestrian samples. The proposed method has been tested on the publicly accessible OSU Color-Thermal dataset, and the obtained results demonstrate the effectiveness of the proposed method when compared to the recent state of the art pedestrian detection approaches. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.

关键词： Intelligent systems

来源：评论

学校读者我要写书评

暂无评论

Predictive Analysis of Sarajevo's AQI using Machine Learning Models for Varied Data Granularity and Prediction Windows 47

Predictive Analysis of Sarajevo's AQI using Machine Learning...

引用

47th ICT and Electronics Convention, MIPRO 2024

作者： Zolota, Emina Hasić, Vahidin Mević, Amina Delić, Amra Krivić, Senka University Of Sarajevo Faculty Of Electrical Engineering Department For Computer Science And Informatics Sarajevo Bosnia and Herzegovina

ISBN: (纸本)9798350382495

This study scrutinizes five years of Sarajevo's Air Quality Index (AQI) data using diverse machine learning models - Fourier autoregressive integrated moving average (Fourier ARIMA), Prophet, and Long short-term memory (LSTM) - to forecast AQI levels. Focusing on various prediction frames, we evaluate model performances and identify optimal strategies for different temporal granularities. Our research unveils subtle insights into each model's efficacy, shedding light on their strengths and limitations in predicting AQI across varied timeframes. This research presents a robust framework for automatic optimization of AQI predictions, emphasizing the influence of temporal granularity on prediction accuracy, automatically selecting the most efficient models and parameters. These insights hold significant implications for data-driven decision-making in urban air quality control, paving the way for proactive and targeted interventions to improve air quality in Sarajevo and similar urban environments. © 2024 IEEE.

关键词： Air quality

来源：评论

学校读者我要写书评

暂无评论

Classification of Functions Using Machine Learning

引用

International Journal of Unconventional Computing 2023年第2-3期18卷 217-247页

作者： Lukac, Martin Yessenbayeva, Aigerim Lewis, Michael Podlaski, Krzysztof Nazarbayev University Department of Computer Science Astana01000 Kazakhstan University of Lodz Faculty of Physics and Applied Informatics Lodz Poland

We study the approximation of function classification by machine learn-ing. We first determine all the NPN and NPNP classes of three-bit reversible and one to five-bits irreversible functions. Then we apply machine learning to learn the labels for each set of functions. We then search the parameter space for several optimization criteria, such as the size of the neural network, the architecture, and learning parameters that would allow the network under training to learn the NPN classification problem with up to 100% accuracy. We also determine if, for n-bit functions, a classification can be performed using two smaller net-works, each classifying n-1-bit functions. Additionally, we discuss the logic solvability of all four-bit functions using the game of Nono-grams. We determine a relation between the learnability of the classification and logic solvability. We determined that while there is a set of functions that can be classified by representing a function as a prod-uct of two subfunctions, a more complex model is needed to classify a n-bit NPN functions as a function of n-1-bit NPN functions. In addi-tion, we determine that approximate classification is always possible, with certain NPN classes being perfectly classified while others are not. Moreover, we can observe a direct relationship between the learnability of certain classes and the speed of finding a solution for these classes when used in a logic game. Finally, we show that the approximation of the classification using machine learning can replace standard database-based classification. The advantages of machine learning methods are end-to-end learning, the lack of circuit design requirements for classifi-cation, and the ability to easily learn the class distinction problem. © 2023 Old City Publishing, Inc.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Sentiment Analysis Using K-NN Algorithm Through Random Search Cross Validation Approach 4

Sentiment Analysis Using K-NN Algorithm Through Random Searc...

引用

4th International Conference of Science and Information Technology in Smart Administration, ICSINTESA 2024

作者： Senda, Tito Afwi Sitompul, Opim Salim Amalia, Amalia Universitas Sumatera Utara Master of Informatics Faculty of Computer Science and Information Technology Medan Indonesia

ISBN: (纸本)9798350376111

This research discusses the application of the KNearest Neighbor (K-NN) algorithm in sentiment analysis, specifically in classifying positive, negative, or neutral sentiments from ChatGPT user opinions collected from user reviews on the Play Store. This research focuses on optimizing the key parameters of the K-NN algorithm using the Random Search cross-validation (RSCV) method. The evaluation results showed a good performance improvement after applying RSCV. Previously, the model showed unsatisfactory performance but improved significantly after optimization. Overall, RSCV successfully improved the average performance of the model for sentiment analysis. © 2024 IEEE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Fine-Tuning GPT-4o-Mini for Programming Questions Generation

Fine-Tuning GPT-4o-Mini for Programming Questions Generation

引用

2024 International Conference on computer and Applications, ICCA 2024

作者： Hesham, Alaa Hamdy, Abeer The British University in Egypt Faculty of Informatics and Computer Science Department of Software Engineering Cairo Egypt

ISBN: (纸本)9798350367560

Creating programming questions that are both meaningful and educationally relevant is a critical task in computer science education. This paper introduces a fine-tuned GPT4o-mini model (C2Q). It is designed to generate meaningful questions by leveraging semantic feature extraction and well- crafted prompts. The approach addresses the limitations of traditional generative models, offering a deeper understanding of programming code and producing questions that are precise, diverse, and relevant to a given code snippets. The proposed framework incorporates essential code elements, such as control structures and method attributes, to generate questions that align with programming concepts. Evaluation metrics used were BLEU, ROUGE-1, and ROUGE-L to evaluate the model's performance. The findings reveal that the model achieves better structural coherence and conceptual relevance while focusing on contextual understanding over exact term matching. This work highlights the potential of the proposed approach to advance teaching and assessment methods in computer science. © 2024 IEEE.

关键词： Program processors

来源：评论

学校读者我要写书评

暂无评论

The Influence of ChatGPT on Student Learning and Academic Performance 5

The Influence of ChatGPT on Student Learning and Academic Pe...

引用

5th International Conference on computer and Applications, ICCA 2023

作者： Ayman, Shehab Eldeen El-Seoud, Samir A. Nagaty, Khaled Karam, Omar H. The British University In Egypt Faculty of Informatics and Computer Science Cairo Egypt

ISBN: (纸本)9798350303254

This study delves into the integration of ChatGPT, an artificial intelligence-driven language model, within undergraduate education. The research scrutinizes the potential advantages, obstacles, and ethical dimensions linked to including ChatGPT in educational practices. It assesses how ChatGPT may impact student engagement, critical thinking, problem-solving abilities, writing proficiency, and the delivery of personalized learning experiences. A survey administered to university faculty members captures their perspectives on ChatGPT usage, encompassing its influence on learning outcomes and academic performance. An internal inquiry at the British University in Egypt found high plagiarism rates in various faculties, with mass media having the highest at 66%, while political science had the lowest at 28%. A survey has been conducted on teaching staff at the university. 75% of teaching staff believes that ChatGPT should be integrated into teaching. However, they are still undecided about whether ChatGPT has affected the teaching and learning process. The study underscores the significance of responsible implementation, faculty training, and continuous assessment to optimize ChatGPT's benefits while adhering to ethical standards in education. In conclusion, this paper illuminates ChatGPT's potential as a valuable tool in undergraduate education, underscoring the imperative of preserving critical thinking skills and human interaction in the learning journey. © 2023 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Modeling Antenna Array with Vibrator Antennas for Problem of Dead on Arrival

Modeling Antenna Array with Vibrator Antennas for Problem of...

引用

2023 International Conference on Industrial Engineering, Applications and Manufacturing, ICIEAM 2023

作者： Preobrazhenskiy, Yu. Avetisyan, T. Preobrazhenskiy, A. Voronezh Institute of High Technologies Faculty of Informatics Voronezh Russia Voronezh Institute of High Technologies Faculty of Informatics Computer Voronezh Russia

ISBN: (纸本)9781665475952

The paper is devoted to the problem of dead on arrival (DoA) sources of radiation of electromagnetic waves. It is proposed to use the antenna array to solve this problem, which contains logoperiodic antennas with vibrators. The solution to an integral equation was carried out to determine the electromagnetic field, which received by the antenna structure. When determining currents on the antenna, the method of collocation was used. To determine the DoA of the sought sources, it was proposed to use the technique, which consisted of two stages. At the first stage, the criterion of the maximum of the magnitudes of the amplitudes of signals at the outputs of the logoperiodic antennas with vibrators were used. It was found that the phase difference from the angle of the electromagnetic wave arrival and on frequency obeys practically linear laws. In the second stage, the direction-finding angle can be estimated for one of the angles of the electromagnetic wave. At the same time, an approximating dependence of the difference of DoA from the angle of the wave is used. The error of the resulting angle of the direction finding does not exceed 1°. © 2023 IEEE.

关键词： Electromagnetic waves

来源：评论

学校读者我要写书评

暂无评论

Job Resumes Recommendation using Integration of Fuzzy Discernibility Matrix Feature Selection and Convolutional Neural Network Multi-label Text Classification

Informatica (Slovenia)

引用

Informatica (Slovenia) 2025年第13期49卷 49-58页

作者： Hayami, Regiolina Yusoff, Nooraini Daud, Kauthar Mohd Fatma, Yulia Department of Informatics Engineering Faculty Computer Science Universitas Muhammadiyah Riau Indonesia Faculty of Data Science and Computing Universiti Malaysia Kelantan 16100 Malaysia Center for Artificial Intelligence Technology Faculty of Information Science and Technology Universiti Kebangsaan Malaysia 43600 Malaysia

The manual analysis of job resumes poses specific challenges, including the time-intensive process and the high likelihood of human error, emphasizing the need for automation in content-based recommendations. Recent advancements in deep learning, particularly Convolutional Neural Networks (CNN) for Multi-label Text Classification (MLTC), offer a promising solution for addressing these challenges through artificial intelligence. While CNN is renowned for its robust feature extraction capabilities, it faces specific challenges such as managing high-dimensional data and poor data interpretability. To address these limitations, this study employs the Fuzzy Discernibility Matrix (FDM) feature selection technique to determine the relevance of skills for each job vacancy. FDM assigns weights to the features of each job category and ranks their relevance based on the highest scores, which are then utilized in the MLTC CNN model. The integration of FDM and MLTC CNN serves as the foundation for generating content-based recommendations derived from key features in job resumes. This study produces a content-based job recommendation system that displays the top three job categories, accompanied by explanations of the skills supporting each selected category. These recommendations also consider cosine similarity values between analyzed items and the integrated results of FDM and MLTC CNN. The application of FDM for feature weighting has been proven to enhance multi-label text classification outcomes by providing better insights during the feature selection process. With a recall of 97.26%, precision of 94.81%, and accuracy of 98.58%, the MLTC model integrating FDM feature selection and CNN demonstrates robust performance characteristics in content-based job recommendation tasks. © 2025 Slovene Society Informatika. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Tiny Machine Learning for Underwater Image Enhancement: Pruning and Quantization Approach 5

Tiny Machine Learning for Underwater Image Enhancement: Prun...

引用

5th International Conference on computer and Applications, ICCA 2023

作者： Rejal, Ayah Abo El Pester, Andreas Nagaty, Khaled The British University in Egypt Faculty of Informatics and Computer Science Department of Computer Science Cairo Egypt

ISBN: (纸本)9798350303254

Many people have expressed an interest in underwater image processing in a variety of fields, including underwater vehicle control, archaeology, marine biological studies, etc. Underwater exploration is becoming an increasingly important element of our lives, with applications ranging from underwater marine and creature research to pipeline and communication logistics, touristic and entertainment use. Underwater images suffer from poor visibility, distortion, and poor quality for a variety of causes, including light propagation. The major issue arises when these images must be captured at depths greater than 500 feet and artificial lighting needs to be provided. Efficient algorithms and models were proposed to enhance the images taken underwater. However, these networks challenge efficient real-time deployment and needs significant computational resources and energy costs. In this paper these challenges will be tackled by using different network compression techniques such as pruning and quantization. These techniques will be applied on a model named WGH-net and the resulting in a tiny machine learning model is named tiny WGH-net model. Moreover, we compare the results of both models along with other models proposed by other researchers. Sustainable Development Goals (SDG) will also be discussed along with our contribution to some of the goals. © 2023 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Artificial Bee Colony for Automated Black-Box Testing of RESTful API 1

引用

11th International Conference on Frontiers of Intelligent Computing: Theory and Applications, FICTA 2023

作者： Ahmed, Seif Hamdy, Abeer Faculty of Informatics and Computer Science The British University in Egypt Cairo Egypt

ISBN: (数字)9789819967063

ISBN: (纸本)9789819967056

Recently, RESTful APIs are widely utilized in a variety of web applications;developers utilize RESTful APIs as a blacked-out component in micro-services. Black-box testing for RESTful APIs is essential as neither of the API’s source code nor its compiled binary is always available for public access. A handful number of research studies were conducted for the automatic generation of test suite for RESTful APIs based on black-box testing. However, to our knowledge, none of them considered the test coverage criteria or the test suite optimization. This paper proposed adapting the Artificial Bee Colony (ABC) swarm intelligence algorithm for the automatic generation of test suites for RESTful APIs based on OpenAPI Specification (OAS), while also maximizing the API test coverage (path, operation, parameter, input value, and status code). Experiments were conducted on six APIs that differ from each other in the number of routes, operation types, input values, and how well the API is documented. The experiments showed that the ABC algorithm can generate test suites that achieve high coverage criteria. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Web services

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：