Application Architecture Research Articles

Comparative Study of Adversarial Defenses: Adversarial Training and Regularization in Vision Transformers and CNNs

Transformer-based models are driving a significant revolution in the field of machine learning at the moment. Among these innovations, vision transformers (ViTs) stand out for their application of transformer architectures to vision-related tasks. By demonstrating performance as good, if not better, than traditional convolutional neural networks (CNNs), ViTs have managed to capture considerable interest in the field. This study focuses on the resilience of ViTs and CNNs in the face of adversarial attacks. Such attacks, which introduce noise into the input of machine learning models to produce incorrect outputs, pose significant challenges to the reliability of machine learning models. Our analysis evaluated the adversarial robustness of CNNs and ViTs by using regularization techniques and adversarial training methods. Adversarial training, in particular, represents a traditional approach to boosting defenses against these attacks. Despite its prominent use, our findings reveal that regularization techniques enable vision transformers and, in most cases, CNNs to enhance adversarial defenses more effectively. Through testing datasets like CIFAR-10 and CIFAR-100, we demonstrate that vision transformers, especially when combined with effective regularization strategies, demonstrate adversarial robustness, even without adversarial training. Two main inferences can be drawn from our findings. Firstly, it emphasizes how effectively vision transformers could strengthen artificial intelligence defenses against adversarial attacks. Secondly, it shows how regularization, which requires much fewer computational resources and covers a wide range of adversarial attacks, can be effective for adversarial defenses. Understanding and improving a model’s resilience to adversarial attacks is crucial for developing secure, dependable systems that can handle the complexity of real-world applications as artificial intelligence and machine learning technologies advance.

The next-gen federated search architecture for biomedical knowledge repositories — The LIT-FED-SEARCH engine

The primary objective of LIT-FED-SEARCH software is to develop a user-friendly solution tailored to researchers and scientists. This solution aims to enhance their impact by facilitating the analysis of data from modern, extensive datasets like PubMed and Clinical Trials, alongside real-world evidence. The central concept we offer is a Federated Search Workflow Engine, which has been designed and maintained to accommodate various infrastructure configurations for the convenience of users. In line with this approach, potential users have the flexibility to configure their own computing environment and a set of interesting data repositories, based on their specific requirements and capabilities. This customization can significantly reduce the time and resources invested in research. LIT-FED-SEARCH is constructed with the support of OpenSearch full-text search engine as its heart. This paper offers an overview of the system’s architecture, capabilities, and potential applications in the field of biomedical research.

Automated weld defect segmentation from phased array ultrasonic data based on U-net architecture

Ultrasonic inspection is an environmentally friendly and easily deployable nondestructive testing (NDT) method widely used for defect detection of critical components in the industry. Phased array ultrasonic testing (PAUT) is one of the most advanced ultrasonic inspection methods, which gives volume inspection with increased resolution and coverage, improving inspection efficiency. Because of the weld structure echoes and the abstract nature of ultrasound images, especially facing meters and feature-changing weld joints with different thicknesses and welding methods in shipbuilding, the analysis of the PAUT weld data still relies on experienced rater random inspections. Rating complex welded products process is lengthy, costly, and prone to introduce human error during the manual rating but challenges automatic detection. To automatically segment defects in PAUT data, this work shows a combination of PAUT data of ship weld and three-dimensional (3D) U-net architecture. Combining PAUT imaging principles with welding and scan processes, a PAUT volumetric image dataset, including different thicknesses and scan angles, is established. We pioneered the application of 3D U-net architecture to segment defects in PAUT volume data. We found that U-net architecture with two encoding stages will perform better in segmenting defects in PAUT data, and region-based loss mainly improves the accuracy. Furthermore, a lightweight U-net architecture containing skip-connection and residual blocks is proposed with precision and efficiency improvement. The validation results show that the proposed U-net architecture offers a feasible solution to the problem of segmenting defects from PAUT data with a Dice accuracy of 90.9 %. Segmentation results help to locate and measure defects. This method makes locating and sizing defects in PAUT weld data possible within a fraction of a second.

Factors Impacting the Photostability of a Roll‐to‐Roll Processed PEDOT Derivative for Flexible Electrochromic Devices

AbstractFor the use of polymeric electrochromic devices (ECDs) in architectural or automotive applications, the photostability of the electrochromic (EC) polymer is crucial for long‐term durability. In this study, the photostability of the side‐chain modified poly(3,4‐ethylenedioxythiophene) (PEDOT‐EthC6)thin films deposited by roll‐to‐roll slot‐die coating on indium tin oxide coated polyethylene terephtalate (PET‐ITO) is investigated and characterized by UV–vis and IR spectroscopy. Residual iron salt on the EC polymer layer, necessary for the in situ polymerization, and the variation of storage times between processing steps have no influence on the degradation rate during light exposure. The photostability of the thin films is 30% better in an inert atmosphere than in ambient conditions. Different long‐pass filters are applied to enhance the photostability in the colored and the bleached state. The position of the absorption edge strongly affects the photostability and visual appearance/color of the PEDOT‐EthC6 electrodes. The analysis of the CIE L*a*b* color coordinates reveals that a trade‐off between stability and a desirable color is necessary. To prevent UV degradation of the EC polymer in flexible ECDs, sputtered nickel oxide is chosen as the counter electrode material. The hybrid ECD shows no sign of degradation and loss of electrochromic performance after 350 h of light exposure.

Analysis and Design of Integration Model for API Management and CI/CD at Directorate General of Taxation

The Directorate General of Taxes (DGT) currently utilizes Application Programming Interface (API) to enhance efficiency in tax data exchange with external parties. DGT is facing challenges due to the rising number of published APIs and the increasing connections from external parties to the DGT system, which necessitates a speedy API issuance process. The objective of this research is to assist the Directorate General of Taxation (DGT) in developing an integrated API management system with Continuous Integration/Continuous Deployment (CI/CD). The system design process is conducted using the Standards and Architectures for E-Government Application (SAGA) framework, encompassing Enterprise Viewpoint, Technology Viewpoint, Computational Viewpoint, Information Viewpoint, and Engineering Viewpoint. A qualitative method is employed, including interviews to gain insights into the existing issues. Additionally, information regarding systems and technologies is documented for gap analysis. The results of this analysis are then utilized to design the architecture of the API management system, applications, and technologies. This research yields a model of the API management system integrated with CI/CD at DGT. The model is developed using 3Scale and Jenkins software. Following validation, the API management system at DGT operates effectively with three DGT API systems and three API users.

Future rescue management: Design specifications for a 5G-enabled emergency response information system

Today, access to real-time information from emergency scenes is still limited for emergency medical services, fire departments, and their professionals, also called first responders. Emergency Response Information Systems (ERISs) have recently been discussed in the literature as a potential solution to this problem. Using the Design Science Research (DSR) paradigm, we present a novel 5G-enabled ERIS (5G-ERIS) design that leverages 5G mobile network technologies to offer diverse real-time information. We provide a user-centered examination of design specifications for a 5G-ERIS based on a smart city digital twin. Based on literature and qualitative expert interviews with several first responders in Germany, we derive how emergency medical services and fire departments can improve their decision-making with this 5G-ERIS. Based on existing 5G application architectures, we structure our identified design specifications into four system layers. Our findings provide an essential knowledge base for the successful development, deployment, and long-term use of 5G-ERISs. We stimulate a broader discussion on the design objectives and specifications of 5G-ERISs in theory and practice.

From Quarry to Monument: Considering Mardin Stone (SE, Türkiye) as the Symbol of Architectural and Cultural Heritage

The cultural and architectural significance of stone becomes evident as one traces its journey from quarry to monument. Located in southeastern Türkiye, Mardin City presents an urban landscape characterized by stone-built heritage. This transforms Mardin into an open-air museum where the final stone products are exhibited. In addition to its cultural and historical significance, the present study explores the geological, physical, mechanical, mineralogical, petrographic, and chemical properties of the Mardin Stone through the samples collected from the Artuklu, Midyat and Savur districts of Mardin. The Mardin Stone is represented by two different types of limestone deposited during the Cretaceous-Eocene period, categorized as fossiliferous micritic limestone and micritic limestone. The physicomechanical and thermal characteristics suggest that the samples collected from the Artuklu district exhibit better engineering properties than those from the Midyat and Savur districts. The study mapped ancient and modern quarries, illustrating the significance of the stone extraction and crafting techniques used to produce the Mardin Stone. Additionally, it emphasized the ongoing use of stone in contemporary structures, its vital role in heritage conservation, and its export to various countries, notably for use in the Syriac diaspora churches. The evaluations demonstrate that the Mardin Stone, with its diverse architectural applications and symbolic element of the collective memory, meets the essential standards and criteria of the International Union of Geological Sciences-Heritage Stone Subcommission (IUGS-HSS) as a potential candidate for designation as Heritage Stone.

Diabetic Retinopathy Detection from Retinal Images

Diabetic Retinopathy (DR) is an eye disease that can occur in people with diabetes causing permanent or temporary blindness, where the blood vessels of retina, a layer in rear interior eye and so called light sensitive part effected by high sugar damages nerves. As I mentioned above, DR can present in various ways; Some people with sick retinas experience blurred vision, while others have difficulty seeing colors or eye floaters. At first, DR might cause no symptoms or only mild vision problems. But it can lead to blindness if left undiagnosed and untreated. This study presents a pipeline for processing and analyzing retinal images. The pipeline consists of three primary steps: (i) pre-processing images, (ii) extracting features, and (iii) classifying the results. The initial step involves pre-processing the image using various modifications to improve its quality and standardize it. Gaussian filtering has been demonstrated to work fairly well in this situation for improving the contrast of the photos. Convolutional Neural Network (CNN) based pre-trained machine learning algorithm is used to speed up the diagnosis of the severity of DR using retinal images of patients. Convolution neural networks (CNNs), among the greatest neural network architectures for image processing applications, have been utilized in the second and third stages. Keywords-Gaussian filtering,Convolutional Neural Network,Retinalimages,Diabetic retinopathy,hemorrhage,, and exudates.

Applications of Brain Wave Classification for Controlling an Intelligent Wheelchair

The independence and autonomy of both elderly and disabled people have been a growing concern in today’s society. Therefore, wheelchairs have proven to be fundamental for the movement of these people with physical disabilities in the lower limbs, paralysis, or other type of restrictive diseases. Various adapted sensors can be employed in order to facilitate the wheelchair’s driving experience. This work develops the proof concept of a brain–computer interface (BCI), whose ultimate final goal will be to control an intelligent wheelchair. An event-related (de)synchronization neuro-mechanism will be used, since it corresponds to a synchronization, or desynchronization, in the mu and beta brain rhythms, during the execution, preparation, or imagination of motor actions. Two datasets were used for algorithm development: one from the IV competition of BCIs (A), acquired through twenty-two Ag/AgCl electrodes and encompassing motor imagery of the right and left hands, and feet; and the other (B) was obtained in the laboratory using an Emotiv EPOC headset, also with the same motor imaginary. Regarding feature extraction, several approaches were tested: namely, two versions of the signal’s power spectral density, followed by a filter bank version; the use of respective frequency coefficients; and, finally, two versions of the known method filter bank common spatial pattern (FBCSP). Concerning the results from the second version of FBCSP, dataset A presented an F1-score of 0.797 and a rather low false positive rate of 0.150. Moreover, the correspondent average kappa score reached the value of 0.693, which is in the same order of magnitude as 0.57, obtained by the competition. Regarding dataset B, the average value of the F1-score was 0.651, followed by a kappa score of 0.447, and a false positive rate of 0.471. However, it should be noted that some subjects from this dataset presented F1-scores of 0.747 and 0.911, suggesting that the movement imagery (MI) aptness of different users may influence their performance. In conclusion, it is possible to obtain promising results, using an architecture for a real-time application.

Facial expression recognition (FER) survey: a vision, architectural elements, and future directions

With the cutting-edge advancements in computer vision, facial expression recognition (FER) is an active research area due to its broad practical applications. It has been utilized in various fields, including education, advertising and marketing, entertainment and gaming, health, and transportation. The facial expression recognition-based systems are rapidly evolving due to new challenges, and significant research studies have been conducted on both basic and compound facial expressions of emotions; however, measuring emotions is challenging. Fueled by the recent advancements and challenges to the FER systems, in this article, we have discussed the basics of FER and architectural elements, FER applications and use-cases, FER-based global leading companies, interconnection between FER, Internet of Things (IoT) and Cloud computing, summarize open challenges in-depth to FER technologies, and future directions through utilizing Preferred Reporting Items for Systematic reviews and Meta Analyses Method (PRISMA). In the end, the conclusion and future thoughts are discussed. By overcoming the identified challenges and future directions in this research study, researchers will revolutionize the discipline of facial expression recognition in the future.

Livmats Pavilion: Design and Development of a Novel Building System Based on Natural Fibres Coreless-Wound Structural Components for Applications in Architecture

Robotic coreless filament winding with natural fibres and bio-based resin systems offers potential solutions to address productivity and sustainability challenges in the construction sector. Their use in modular, prefabricated structures enables efficient, controlled, and eco-friendly production, yielding high-quality building components with minimal production waste. Plant fibres like flax provide good strength and stiffness while requiring less non-renewable energy compared to traditional fibres like glass or carbon. This research focuses on the use and implementation of flax fibre for coreless filament wound structural components by describing the conceptualization, design, fabrication, and assembly of a full-scale architectural demonstrator, the livMats Pavilion.

Evaluation of Storage Placement in Computing Continuum for a Robotic Application

This paper analyzes the timing performance of a persistent storage designed for distributed container-based architectures in industrial control applications. The timing performance analysis is conducted using an in-house simulator, which mirrors our testbed specifications. The storage ensures data availability and consistency even in presence of faults. The analysis considers four aspects: 1. placement strategy, 2. design options, 3. data size, and 4. evaluation under faulty conditions. Experimental results considering the timing constraints in industrial applications indicate that the storage solution can meet critical deadlines, particularly under specific failure patterns. Comparison results also reveal that, while the method may underperform current centralized solutions in fault-free conditions, it outperforms the centralized solutions in failure scenario. Moreover, the used evaluation method is applicable for assessing other container-based critical applications with timing constraints that require persistent storage.

The Generalization of Stage-Reduced STPS for Low-Loss Unequal 1 × 4 Phased Array Architecture for 5G IoT Applications

The Generalization of Stage-Reduced STPS for Low-Loss Unequal 1 × 4 Phased Array Architecture for 5G IoT Applications

LDSG-Net: an efficient lightweight convolutional neural network for acute hypotensive episode prediction during ICU hospitalization

Objective. Acute hypotension episode (AHE) is one of the most critical complications in intensive care unit (ICU). A timely and precise AHE prediction system can provide clinicians with sufficient time to respond with proper therapeutic measures, playing a crucial role in saving patients’ lives. Recent studies have focused on utilizing more complex models to improve predictive performance. However, these models are not suitable for clinical application due to limited computing resources for bedside monitors. Approach. To address this challenge, we propose an efficient lightweight dilated shuffle group network. It effectively incorporates shuffling operations into grouped convolutions on the channel and dilated convolutions on the temporal dimension, enhancing global and local feature extraction while reducing computational load. Main results. Our benchmarking experiments on the MIMIC-III and VitalDB datasets, comprising 6036 samples from 1304 patients and 2958 samples from 1047 patients, respectively, demonstrate that our model outperforms other state-of-the-art lightweight CNNs in terms of balancing parameters and computational complexity. Additionally, we discovered that the utilization of multiple physiological signals significantly improves the performance of AHE prediction. External validation on the MIMIC-IV dataset confirmed our findings, with prediction accuracy for AHE 5 min prior reaching 93.04% and 92.04% on the MIMIC-III and VitalDB datasets, respectively, and 89.47% in external verification. Significance. Our study demonstrates the potential of lightweight CNN architectures in clinical applications, providing a promising solution for real-time AHE prediction under resource constraints in ICU settings, thereby marking a significant step forward in improving patient care.

Stateless Blockchain-Based Lightweight Identity Management Architecture for Industrial IoT Applications

Stateless Blockchain-Based Lightweight Identity Management Architecture for Industrial IoT Applications

Eye Strain Expression Classification using Attention Capsule Network for Adapting Screen Vision

Beside the conventional facial expression recognition methods, the research focuses on developing a system for recognizing various eye expressions under different screen conditions. This research deals with the use of Capsule Network (a recent Deep Learning algorithm) to enhance facial expression recognition capabilities and to develop adaptive screen technologies aimed at mitigating digital eye strain. The main objective of this research is to engineer a sophisticated system that employs the capabilities of Capsule Nets to recognize the various expressions that user makes and based on the recognized expression, dynamically modify screen settings, ensuring optimal user visual comfort. The research primarily concentrates on the exploration and application of various Capsule Net architectures designed for the recognition of expressions related to eye strain. The baseline model utilized elementary convolutional layers which feed into subsequent fully connected layers for the task of classification. The model has since been refined by incorporating advanced techniques such as attention mechanisms and more sophisticated network architectures where the classification is done by Capsule Network. Results have demonstrated a modest enhancement in the Capsule Net’s predictive performance, attributed to its superior spatial and hierarchical processing of facial features, in comparison to conventional deep learning approaches. The final model has an accuracy of 82.27%. As a final system the model has been deployed to an application to process frames from video camera in the device and make prediction to prompt the notifications or recommendations.

Analysing and managing risk from third-party OAuth application access

Modern cloud application architectures allow users to dynamically grant third-party apps access to their cloud resources. When looking at a large, anonymised dataset of over 600,000 users and their approvals of over 43,000 applications, we found that organisations often are unaware of the magnitude of the access problems: on average, an organisation will grant 440 unique third-party apps access to Google data and resources; one organisation’s users approved 12,330 unique applications; out of all the approved applications, over 44 per cent have been granted access to either sensitive data or all data on the user’s Google Drive. The use of OAuth 2.0 has resulted in users creating numerous, unknown access paths from external third-party apps to organisations’ data and resources, in a way that is often hidden and unmanaged by organisations’ IT or security departments. In addition, the credentials granted to access users’ cloud data and resources often can be refreshed indefinitely. Security operations face challenges such as minimal to no tracking or management of these paths or credentials, along with immature or incomplete technical controls. This paper discusses an approach to understanding, assessing, analysing and managing third-party OAuth application risk, using anonymised real-world data as relevant examples, to provide prescriptive guidance on reducing risk associated with third-party applications.

Contemporary Architectural Applications of Moroccan Zellij in Urban Spaces

This paper explores the evolving role of Moroccan Zellij in contemporary urban architectural design, highlighting its integration into global projects and its potential as a sustainable design element. Zellij, a traditional Moroccan tile work, is renowned for its intricate patterns and vibrant colors. As urban environments increasingly seek sustainable and culturally rich materials, Zellij has found a new relevance in modern architecture. The paper analyzes how Zellij is being adapted for use in diverse settings, from luxury buildings to public spaces, and discusses the technical and logistical challenges involved in its integration. It also examines the aesthetic and cultural impacts of Zellij on urban spaces, its contribution to energy efficiency and environmental sustainability, and how it fosters a deeper connection between traditional craftsmanship and contemporary design. Finally, the paper forecasts future trends in the use of Zellij and its potential influence on global architectural styles, suggesting that Zellij will continue to enhance urban landscapes while promoting sustainable building practices.

Modern Trends in Green Interior Architecture Applications with Smart Systems

Modern Trends in Green Interior Architecture Applications with Smart Systems

Towards Urban Digital Twins: A Workflow for Procedural Visualization Using Geospatial Data

A key feature for urban digital twins (DTs) is an automatically generated detailed 3D representation of the built and unbuilt environment from aerial imagery, footprints, LiDAR, or a fusion of these. Such 3D models have applications in architecture, civil engineering, urban planning, construction, real estate, Geographical Information Systems (GIS), and many other areas. While the visualization of large-scale data in conjunction with the generated 3D models is often a recurring and resource-intensive task, an automated workflow is complex, requiring many steps to achieve a high-quality visualization. Methods for building reconstruction approaches have come a long way, from previously manual approaches to semi-automatic or automatic approaches. This paper aims to complement existing methods of 3D building generation. First, we present a literature review covering different options for procedural context generation and visualization methods, focusing on workflows and data pipelines. Next, we present a semi-automated workflow that extends the building reconstruction pipeline to include procedural context generation using Python and Unreal Engine. Finally, we propose a workflow for integrating various types of large-scale urban analysis data for visualization. We conclude with a series of challenges faced in achieving such pipelines and the limitations of the current approach. However, the steps for a complete, end-to-end solution involve further developing robust systems for building detection, rooftop recognition, and geometry generation and importing and visualizing data in the same 3D environment, highlighting a need for further research and development in this field.