Applied and Computational Engineering

- The Open Access Proceedings Series for Conferences

Volume Info.

  • Title

    Proceedings of the 4th International Conference on Signal Processing and Machine Learning

    Conference Date

    2024-01-15

    Website

    https://www.confspml.org/

    Notes

     

    ISBN

    978-1-83558-347-0 (Print)

    978-1-83558-348-7 (Online)

    Published Date

    2024-03-25

    Editors

    Marwan Omar, Illinois Institute of Technology

Articles

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241105

    DCGAN-based data augmentation for improving CNN performance in meningioma classification

    Using Magnetic Resonance Imaging (MRI) to scan the brain and using the image to identify whether the patient has a brain tumor or not is the common way doctors use it today. However, as this may potentially add to the workload of healthcare professionals, it becomes crucial to explore methods for automating image identification. One effective algorithm for this purpose is the utilization of a Convolutional Neural Network (CNN) network. However, when applying a CNN network to discern whether an individual has meningioma or not, it becomes evident that the available data may be limited. Meningioma is relatively uncommon, and not all associated images have been made accessible for analysis. The shortage of original samples makes it hard to train the CNN network and has relatively low accuracy. In this case, this study tries to use DCGAN to generate more images based on the original sample. By comparing the accuracy and f1 score of the CNN network, this study finds that implementing images has improved the performance of the CNN network. By implementing the images, the DCGAN generates, the accuracy for the same CNN network to identify whether the images have meningioma or not is increased from 93.53 percent to 97.75 percent. The f1 score also increased from about 0.9187 to about 0.9738.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241147

    QR code resolution improvement based on Super-Resolution Generative Adversarial Network

    QR codes have become an integral part of our daily routines, simplifying tasks ranging from accessing websites to making payments. However, the quality of QR codes, especially their resolution, can significantly impact their functionality. Low-resolution QR codes may lead to misinterpretation during scanning and even decoding failures. To address this issue, researchers have explored various techniques to enhance the resolution of QR codes. Traditional image processing methods, such as interpolation and filtering, have been used in the past for resolution enhancement. However, these methods often result in overly blurry images with poor perceptual quality. Conversely, solutions based on Convolutional Neural Networks (CNNs) can introduce clarity but may compromise the sharpness of image edges. This paper presents an effective approach to improve QR code resolution using a Super-Resolution Generative Adversarial Network (SRGAN). The results are impressive, with SRGAN achieving a Peak Signal-to-Noise Ratio (PSNR) of 30.06, significantly outperforming the 17.48 achieved by the SRCNN method. Additionally, in terms of Structural Similarity Index (SSIM), SRGAN reaches 0.936, surpassing SRCNN's 0.473. These metrics demonstrate that SRGAN is highly effective in enhancing the resolution of QR codes, ensuring better scan accuracy and overall functionality in practical applications.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241148

    Medical image super-resolution reconstruction: A comprehensive investigation of Generative Adversarial Networks

    Medical images play a crucial role in modern healthcare diagnostics and treatment. However, many medical images suffer from limitations in resolution, potentially impeding a comprehensive understanding of a patient's condition by healthcare professionals. This comprehensive review delves into the applications of Generative Adversarial Networks (GANs) in medical image super-resolution reconstruction to address this challenge. In the Methods section, this paper first focused on the direction of medical image classification, including cell classification of histopathological images and synthetic data enhancement using GANs to improve liver lesion classification. Subsequently, this paper focused on the direction of medical image segmentation, looking into the use of Structure-Corrected Adversarial Networks (SCAN) for organ segmentation in chest radiographs and Deep Adversarial Networks for biomedical image segmentation using unannotated images. In the Applications and Discussion section, this paper thoroughly examined the current progress of GANs in telemedicine diagnosis and disease state generation and prediction. This paper emphasized the significant potential of GAN technology in telemedicine while outlining the current constraints and challenges. Furthermore, this paper highlighted the prospects of GANs in medical image super-resolution reconstruction and how they affect the discipline of medical imaging. This comprehensive review consolidates the latest research findings on GANs in medical image super-resolution reconstruction, underscoring their importance in the realm of healthcare. By critically analysing existing literature, this paper provides valuable insights for medical image analysts are researchers while inspiring future research directions and innovations.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241152

    Enhanced DeblurGAN: An advanced combinatorial model for motion blur removal in low-light photography

    This article aims to address the challenge of eliminating low-light motion blur, a problem that lacks effective solutions, despite being crucial in various application scenarios. For instance, it can help in the identification of moving individuals or license plates during nocturnal surveillance, filming running videos after dark, and managing animals in rural areas at night. These examples represent commonplace and significant scenarios. These are all important domains, but few approaches are effective at handling such specific cases simultaneously. This paper utilizes a fusion model to increase the brightness of an image while preserving the photographic details. The motion blur is subsequently eliminated from the brightness-enhanced image. This results in the enhancement of image details and the removal of motion blur. Comparing the model proposed in this paper with the commonly used Deblur model, it becomes apparent that the new model effectively enhances brightness in low-light motion blur while preserving image details and reducing much of the blur. This implies that the model is more versatile, as it can be used not only for images but also for low-light videos.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241153

    Performance analysis of three recommendation algorithms on Amazon datasets

    Recommendation systems play a crucial role in enhancing user satisfaction and driving sales for businesses. They are essential in today’s marketplaces, as they are able to suggest products and services that may interest a particular individual based on their past purchases. In this paper, empirical research is conducted on three distinct subsets of the Amazon dataset, namely Sports & Outdoors, Movies & TV, and Video Games, to comprehensively evaluate the performance of three distinct recommendation methods based on deep learning algorithms. These methods include the dot product method, (an updated version of the singular value decomposition algorithm), the neural network method, and the neural collaborative filtering model with natural language processing method. The results of this study reveal that deep learning-based recommendation systems can achieve more accurate results compared to traditional recommendation systems for three types of products. The implementation of these methods on Amazon’s dataset can help improve sales by correctly identifying customers’ interests and suggesting relevant items.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241155

    An improved focal loss for imbalanced network traffic classification

    With the continued progress of the Internet, network security has become an increasingly significant issue that requires constant attention and research. Network traffic classification is a key technology used to detect and prevent malicious network activity, and it has accordingly received increasing attention and research. However, datasets related to malicious network traffic classification often have imbalanced characteristics. In conventional traffic classification problems with multiple categories, the sample size characteristics of small categories are often overlooked. To address this issue, the focal loss function was proposed, which focuses on small samples by modulating the trade-off between the positive and negative samples through two hyperparameters α and γ. This article uses convolutional neural networks (CNN) to tackle traffic classification problem and explore the optimal values of α parameters in this application scenario. Additionally, this work proposed a novel weight allocation formula to replace α, which allowed small class traffic to obtain higher accuracy.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241158

    Empowering safe and secure autonomy: Federated learning in the era of autonomous driving

    Artificial Intelligence (AI) has a significant impact on empowering autonomous driving systems to perceive and interpret the environment effectively. However, ensuring data privacy and security in autonomous driving systems is a critical challenge. To surmount these hurdles, federated learning has emerged as an effective strategy. Federated learning is a decentralized machine learning approach that facilitates the cooperative training of models across a diverse set of connected devices, enabling them to collectively learn and improve their performance, while preserving data privacy. This approach eliminates the necessity of sharing raw data and only involves sharing model updates with a central aggregator, thereby ensuring privacy and minimizing data exposure. This paper examines the implementation of federated learning in autonomous driving. It explores the principles of federated learning, including decentralized training, local model updates, model aggregation, privacy preservation, iterative learning, and heterogeneity handling. Two specific approaches, Deep Federated Learning (DFL) and End-to-End Federated Learning, are discussed, highlighting their benefits in enhancing privacy and maintaining prediction accuracy. The paper also discusses the applications of federated learning in communication and control aspects of autonomous driving. It emphasizes the scalability, adaptability, edge computing, real-time learning, federated transfer learning, and privacy-preserving data sharing as potential future prospects for federated learning in autonomous driving. Overall, federated learning offers a unique opportunity to address privacy concerns in autonomous driving systems while harnessing the collective intelligence of a fleet of vehicles. It has the potential to revolutionize the field and contribute to the development of safe and secure autonomous driving technologies.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241159

    Revolutionizing comic coloring: Federated learning-based neural network for efficiency and privacy

    The conventional method of coloring comics can be quite arduous and time-consuming, particularly within the realm of animation, where each frame necessitates individual coloring. Over the past few years, the advent of deep learning technology has introduced fresh prospects for comic coloring. Nonetheless, it's worth noting that deep learning typically demands a substantial volume of data to effectively train models, giving rise to legitimate concerns regarding data privacy. This study proposes an innovative approach that applies federated learning to comic coloring to improve efficiency and address privacy issues. In this research, an existing neural network European Conference on Computer Vision (ECCV16) model was employed using federated learning to partition and train the model on segmented databases. To better quantify the color differences between the original images and the colored images, this study designed a custom loss function. Experimental results demonstrate that this approach has been effective in comic coloring, demonstrating the feasibility of applying federated learning to the task of comic coloring.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241162

    Style transfer for converting images into Chinese landscape painting based on CycleGAN

    Chinese landscape painting is a prevalent and unique mode of artistic expression within traditional Chinese art. It boasts intricate techniques and demands a relatively high level of artistic skill. Recent advancements in artificial intelligence have ushered in the era of image style transfer techniques, making it feasible to convert landscape photographs into stunning Chinese landscape paintings. In this study, the author has developed an image translation model that enables mutual transformation between images of landscapes and Chinese landscape paintings. This technology significantly reduces the difficulty of creating Chinese landscape paintings, allowing more ordinary people in China to experience and appreciate the joy brought by traditional Chinese art. Experimental results indicate that the style transfer model based on CycleGAN has achieved significant success in this scenario. The generated artworks successfully integrate the style of Chinese landscape painting into the original images without altering the original composition and details. As a result, the original photos gain a certain level of artistic value. Additionally, this study innovatively explores the goal of reverse restoration of Chinese landscape paintings into images, highlighting both the similarities and differences with the current research, thus laying the foundation for future studies.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241165

    Research on the application of machine learning in business analytics: Cases of Amazon and eBay

    With the rapid development of the Internet and the rise of e-commerce, commercial enterprises are faced with a large amount of data and a complex market environment. In this situation, machine learning, as a powerful tool, is widely used in the field of business analysis. In this dissertation, we take Amazon and eBay as examples to study the application of machine learning in the company's business analytics, focusing on its role in market prediction, customer behavior analysis and operation optimization. By analyzing the relevant cases, we find that machine learning technology plays an important role in helping companies make more accurate decisions and improve efficiency. Studying the application of Amazon machine learning in business analytics can promote in-depth research on the application of machine learning in business in academia, and promote the application and development of machine learning technology in other business scenarios. Overall, the application of machine learning in business analytics can help companies understand customer behavior, optimize operations, and improve sales results. However, there are still some challenges, such as data quality, algorithm selection and privacy protection. Therefore, further research and innovation are necessary to advance the development of machine learning applications in business analytics.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241166

    Research on method of the data classification with an example

    The huge data from real life is a problem amount the researcher. So, it is important to let us divide them in out project. This article wants to talk about the way of data classification. The article using different method to shows how these methods works, start from the simple word dividing and a little complex method which using the root to be the key word. Then the article shows the complex PCA or k-mean dividing which is always appear with the mechanize learning. Then the article using a true example to shows, how this different way of dividing the data can work for different situation. This article can truly increase the speed of the researcher ‘work and increasing the accuracy of the data analysis by letting the data to be more methodical.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241172

    Credit card transaction fraud detection based on machine learning

    The rise of e-commerce payment systems has been swift due to our society's rapid advancement. However, the transparency and vulnerability of the internet have opened new avenues for illicit access, resulting in a significant surge in financial fraud cases, particularly credit card transaction fraud. Detecting and mitigating such incidents has become crucial. One branch of artificial intelligence, machine learning, offers a potent solution. It operates by utilizing various algorithms and models that rely on established patterns and reasoning without explicit instructions. By processing vast amounts of historical data, machine learning models can identify underlying data relationships, allowing them to make accurate predictions based on input data. Therefore, it emerges as a highly effective method for credit card transaction fraud detection. This paper reviews the research methods for credit card fraud, introduces credit card transaction fraud detection data sets, and outlines machine learning algorithms and models related to credit card fraud detection. It also considers the future prospects for machine learning development and possible challenges.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241174

    TabTranSELU: A transformer adaptation for solving tabular data

    Tabular data are most prevalent datasets in real world, yet the integration of deep learning algorithms in tabular data often garners less attention despite their widespread utilization in other field. This phenomenon could be attributed to the dominance of the classical algorithms in their simplicity and interpretability, and the superior performance of the gradient boosting tree models in tabular data. In this paper, a simple yet affective adaptation of the Transformer architecture tailored specifically for tabular data is presented, not only achieving good performance but also retains a high degree of explain ability. The model encodes both continuous and categorical features, alongside their respective names, and feed them into an enhanced Transformer structure enriched with Scaled Exponential Linear Unit activation. Through rigorous experimentation, our model not only outperforms classical algorithms and similar Transformer-based counterparts, but also are comparable to the performance of gradient boosting tree models.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241178

    Comparative analysis of different model depths on convolutional neural network for handwritten digit recognition

    In the rapidly advancing field of artificial intelligence, Convolutional Neural Networks (CNNs), as a representative method, have risen to prominence as a pivotal instrument for handling visual data. However, despite their widespread use, the impact of CNN depth on performance remains under-explored. This study delves into this aspect, evaluating the performance of CNN architectures with different depths - two-layer, four-layer, and five-layer - on the MNIST dataset, a version from the National Institute of Standards and Technology, a well-known benchmark dataset for handwritten digit recognition. Experimental results reveal that the four-layer model achieved the highest average accuracy of 99.76%, while the five-layer model, despite its additional complexity, only slightly trailed behind with a 99.73% accuracy rate. However, the five-layer model required a significantly longer training time. In conclusion, while deeper networks can increase accuracy, they can also introduce computational inefficiencies without significant gains in performance. This research provides a better understanding of CNN depth, guiding optimal model selection for image classification tasks.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241183

    Data statistical analysis on Amazon e-commerce platform for recommender system

    Recommendation systems are a crucial element in engaging users and maintaining their engagement with e-commerce platforms. By recommending products or services that are likely to be relevant to each user's interests and preferences, the system can help maintain user interest and encourage them to spend more time on the platform. This work analyzes the principles of several off-the-shelf recommendation models, including the collaborative filtering model, the singular value decomposition model, and the rating-based collaborative filtering model. These models play a crucial role in the field of recommendation systems for e-commerce. To gain further insights from the comments of various merchandise items, sentiment analysis techniques and word cloud analysis are applied. Evaluation of these results demonstrates the critical role of recommender systems in shaping the future landscape of e-commerce. Sentiment analysis allows us to identify patterns in user feedback and understand how different factors influence user satisfaction with products or services. Word cloud analysis provides a visual representation of the most frequently mentioned features or keywords in the comments, allowing us to identify trends and patterns in user behavior. By combining these techniques with traditional recommendation models, more accurate and personalized recommendations could be made that better meet user needs and enhance their shopping experience on e-commerce platforms.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241187

    AI in education: Enhancing learning experiences and student outcomes

    This research article makes an attempt to investigate the potential of Artificial Intelligence (AI) in enhancing the learning experiences, as well as student outcomes. As a result, it has developed a study that will be able to understand how the different AI tools, including machine learning, data learning, virtual reality (VR) and augmented reality (AR), automation, and so forth can be used to develop learning experiences and outcomes. Subsequently, a case study involving a mathematics classroom was used to collect data and confirm whether indeed AI led to improved learning experiences and study outcomes. The study confirms that AI resulted in positive outcome with positive performance measure sin academic performance, motivation and engagement, learning progression, and so forth.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241188

    Comparative analysis of strategies of knowledge distillation on BERT for text matching

    Large language model is a highly effective and promising language model technology that can improve the performance and robustness of natural language processing tasks. As a representative work, Bidirectional Encoder Representations from Transformers (BERT) has excellent performance in various natural language processing tasks. This model is pre-trained on large-scale language dataset and has gained attention from all walks of life since its introduction. However, its huge number of participants and scale make its performance in mobile very limited. As an effective technique to compress neural network, knowledge distillation can obtain a lightweight model with smaller parameters without losing too much model performance. Therefore, distillation of BERT models has been started, aiming at obtaining lightweight BERT models. In this paper, we will introduce several common BERT distillation models and analyse their model architecture, distillation process, and finally the compression efficiency and model effectiveness. It is concluded that the process of increasing recompression efficiency is often accompanied by decreasing model effectiveness.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241189

    Machine learning algorithms in the development of recommender systems

    With the continuous development and widespread use of the Internet, individuals often find themselves overwhelmed by the vast amount of information available. To address this issue and better cater to users' personalized needs, recommender systems have emerged. A recommender system is a technology that provides users with customized content or services based on their preferences and interests. Its primary goal is to predict user behavior as accurately as possible in order to recommend relevant items. Machine learning algorithms play a crucial role in the functioning of recommender systems. These algorithms not only automatically analyze and process large volumes of data but also extract valuable features from historical data, enabling accurate predictions for unknown data. The rapid progress of deep learning technology has significantly enhanced the accuracy and efficiency of recommender systems. Deep learning enables recommender systems to better grasp user behavior, interests, and preferences, leading to more precise predictions of user needs and behaviors. Furthermore, deep learning can handle unstructured data such as images, audio, and text, extracting valuable features to improve recommendations. In the future, as machine learning and artificial intelligence technologies continue to advance and gain popularity, recommender systems will find broader applications. Apart from the traditional e-commerce sector, recommender systems can also be applied to social networks, news media, online education, and other fields. Simultaneously, in the research and implementation of recommender systems, due attention must be given to user privacy protection and data security. Only by ensuring user privacy and data security can recommender systems effectively meet user needs and gain wider acceptance.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241192

    Ancient Chinese painting style transfer based on CycleGAN

    Style transfer aims to alter the visual aesthetic of images by giving them a different artistic style. With the rapid advancement of deep learning, style transfer tasks have made significant progress, introducing new perspectives and innovative potentials within the realm of image processing. This study seeks to explore style transfer methods based on Cycle-Consistent Adversarial Networks (CycleGAN), enabling contemporary landscape photographs to take on the form of ancient Chinese paintings. This endeavor opens up fresh possibilities for artistic creation, image editing and design applications. The research encompasses an exposition of the process involved in constructing the CycleGAN model, alongside presenting research findings. Furthermore, it delves into the discussion of crucial techniques employed during the model training process, specifically the utilization of cycle consistency loss in configuring the loss functions. Lastly, this study ventures into future research directions, including strategies for further enhancing the performance and expanding the application scope of this style transfer model.

  • Open Access | Article 2024-03-25 Doi: 10.54254/2755-2721/51/20241200

    An examination of non-wearable methods for fall detection

    Amidst the emergence of an aging global population, elderly healthcare has evolved into a pressing societal concern. The World Health Organization reports that approximately 646,000 individuals across the globe succumb to fatal falls annually. Notably, those aged over 60 experience the highest mortality rate, representing the most significant portion of these deaths. Given these stark statistics, there has been a surge of interest in researching fall detection methodologies. This article delves into non-wearable fall detection techniques, emphasizing their various classifications and applications. Initially, we provide an overview of distinct categories within the non-wearable fall detection landscape. Following this, each method is elaborated upon, elucidating its mechanisms and practical applications. To complement this, a brief discourse on prevalent machine learning algorithms employed in fall detection is presented. In culmination, a comparative analysis of each technique is provided, highlighting their respective merits and limitations. The objective is to furnish readers with a holistic perspective on the current state of non-wearable fall detection. As we gaze towards the horizon, it becomes evident that advancing this domain is paramount to safeguarding our elderly and reducing preventable fatalities.

Copyright © 2023 EWA Publishing. Unless Otherwise Stated