Software analytics (SA) is frequently proposed as a tool to support practitioners in software engineering (SE) tasks. We have observed that several secondary studies on SA have been published. Some of these studies have overlapping aims and some have even been published in the same calendar year. This presents an opportunity to analyze the congruence or divergence of the conclusions in these studies. Such an analysis can help identify broader generalizations beyond any of the individual secondary studies. We identified five secondary studies on the use of SA for SE. These secondary studies cover primary research from 2000 to 2021. Despite the overlapping objectives and search time frames of these secondary studies, there is negligible overlap of primary studies between these secondary studies. Thus, each of them provides an isolated view, and together, they provide a fragmented view, i.e., there is no “common picture” of the area. Thus, we conclude that an overview of the literature identified by these secondary studies would be useful in providing a more comprehensive overview of the topic.

Defect inspection of the surface in ultrasonically welded wire terminations is an important inspection procedure to ensure welding quality. However, the detection task of ultrasonic welding defects based on deep learning still faces the challenges of low detection accuracy and slow inference speed. Therefore, to solve the above problems, we propose a fast and effective lightweight detection model based on You Only Look Once v8 (YOLOv8n), named LightYOLO. Specifically, first, to achieve fast feature extraction, a Two-Convolution module with FasterNet block and Efficient multi-scale attention (CTFE) structures is introduced in the backbone network. Secondly, Group-Shuffle Convolution (GSConv) is used to construct the feature fusion structure of the neck, which enhances the fusion efficiency of multi-level features. Finally, an auxiliary head training method is introduced to extract shallow details of the network. To verify the effectiveness of the proposed method, we constructed a surface defect data set of ultrasonic welding wire terminals and conducted a series of experiments. The results of experiments show that the precision of LightYOLO is 93.4%, which is 3.5% higher than YOLOv8n(89.9%). In addition, the model size was reduced to 1/2 of the baseline model. LightYOLO shows the potential for rapid detection on edge computing devices. The source code and dataset for our project is accessible at https://github.com/JianshuXu/LightYOLO. © 2024 The Authors

In smart cities, Closed-Circuit Television (CCTV) cameras are crucial for enhancing law enforcement and public safety. Since CCTV cameras include both state-controlled and private-sector units in large numbers, it poses a considerable challenge to manage them. Automating the calculation of their coverage areas enables law enforcement and city planners to efficiently adapt surveillance strategies to the evolving needs of urban safety and dynamics.This paper proposes a prototype that automates the digitization of a vast number of camera coverage areas, such as for a large city. Given the positions for each camera, its sector width, and length-of-view, the prototype automatically identifies each camera's capture area represented as a polygon of positions. The prototype's generated capture areas are validated to the ground truth representing the true capture areas for 51 cameras (in the city of Malmö, Sweden) provided by the Swedish law enforcement agencies. © 2024 IEEE.

Plazas are an essential pillar of public life in our cities. Historically, they have been seen as public fora, hosting public events that fostered trade, interaction, and debate. However, with the rise of modern urbanism, city planners considered them as part of a larger strategic development scheme overlooking their social importance. As a result, plazas have lost their function and value. In recent years, awareness has risen of the need to re‐activate these public spaces to strive for social inclusion and urban resilience. Geometric and urban features of plazas and their surroundings often suggest what kinds of usage the public can make of them. In this project, we explore the application of machine learning to predict the suitability of events in public spaces, aiming to enhance urban plaza design. Learning from traditional urbanism indicators, we consider factors associated with the features of the public space, such as the number of people and the high degree of comfort, which are evolved from three subcategories: external factors, geometric shape, and design factors. We acknowledge that the predictive capability of our model is constrained by a relatively small dataset, comprising 15 real plazas in Madrid augmented digitally to 2025 fictional scenarios through self‐organising maps. The article details the methods to quantify and enumerate quantitative urban features. With a categorical target variable, a classification model is trained to predict the type of event in the urban space. The model is then evaluated locally in Grasshopper by visualising a parametric verified geometry and deploying the model on other existing plazas worldwide regarding geographical proximity to Madrid, where to share or not the same cultural and environmental conditions. Despite these limitations, our findings offer valuable insights into the potential of machine learning in urban planning, suggesting pathways for future research to expand upon this foundational study. © 2025 by the author(s).

Context: As the diversity and complexity of regulations affecting Software-Intensive Products and Services (SIPS) is increasing, software engineers need to address the growing regulatory scrutiny. We argue that, as with any other non-negotiable requirements, SIPS compliance should be addressed early in SIPS engineering—i.e., during requirements engineering (RE).

Objectives: In the conditions of the expanding regulatory landscape, existing research offers scattered insights into regulatory compliance of SIPS. This study addresses the pressing need for a structured overview of the state of the art in software RE and its contribution to regulatory compliance of SIPS.

Method: We conducted a systematic mapping study to provide an overview of the current state of research regarding challenges, principles, and practices for regulatory compliance of SIPS related to RE. We focused on the role of RE and its contribution to other SIPS lifecycle process areas. We retrieved 6914 studies published from 2017 (January 1) until 2023 (December 31) from four academic databases, which we filtered down to 280 relevant primary studies.

Results: We identified and categorized the RE-related challenges in regulatory compliance of SIPS and their potential connection to six types of principles and practices addressing challenges. We found that about 13.6% of the primary studies considered the involvement of both software engineers and legal experts in developing principles and practices. About 20.7% of primary studies considered RE in connection to other process areas. Most primary studies focused on a few popular regulation fields (privacy, quality) and application domains (healthcare, software development, avionics). Our results suggest that there can be differences in terms of challenges and involvement of stakeholders across different fields of regulation.

Conclusion: Our findings highlight the need for an in-depth investigation of stakeholders’ roles, relationships between process areas, and specific challenges for distinct regulatory fields to guide research and practice. 

It is commonly accepted that the quality of requirements specifications impacts subsequent software engineering activities. However, we still lack empirical evidence to support organizations in deciding whether their requirements are good enough or impede subsequent activities. We aim to contribute empirical evidence to the effect that requirements quality defects have on a software engineering activity that depends on this requirement. We conduct a controlled experiment in which 25 participants from industry and university generate domain models from four natural language requirements containing different quality defects. We evaluate the resulting models using both frequentist and Bayesian data analysis. Contrary to our expectations, our results show that the use of passive voice only has a minor impact on the resulting domain models. The use of ambiguous pronouns, however, shows a strong effect on various properties of the resulting domain models. Most notably, ambiguous pronouns lead to incorrect associations in domain models. Despite being equally advised against by literature and frequentist methods, the Bayesian data analysis shows that the two investigated quality defects have vastly different impacts on software engineering activities and, hence, deserve different levels of attention. Our employed method can be further utilized by researchers to improve reliable, detailed empirical evidence on requirements quality. © The Author(s) 2024.

Recently, deep learning algorithms and software-defined networking technologies enabled traffic management to be more controllable in IP networking and mobile Internet to yield quality services to subscribers. Quality of service (QoS) needs more effort to optimize QoS performance. More specifically, elephant flow management is a critical task that needs further research since its heavy hit behavior leads to high CPU utilization, packet drops, high latency, packet overflow, and network congestion problems. For this purpose, we focused on elephant flow management since elephant flows are big flows that hinder good service delivery (QoS) on demand. Hence, elephant flow detection and early prediction techniques optimize QoS. In this regard, we employed DNN and CNN deep learning models to detect elephant flows, and the random forest model predicts elephant flows in the SDN. As a result of our experiments, the findings reveal that deep learning algorithms within the Ryu controller significantly outperform in detecting and predicting performance in order to yield good network throughput. This solution proves to be significant for QoS optimization in data centers.

Cartographic generalization aims at decreasing map or database detail. On one hand, its goal is taking into account map purpose, user constraints and needs, on the other hand maintaining and highlighting characteristic spatial patterns. One of the main challenges in the research concerning cartographic generalization is the evaluation of its results. While previous studies have exclusively concentrated on quantitative evaluation of cartographic generalization results, we complement these studies by considering both quantitative and qualitative evaluation with the map designers and map users. In this pilot study, six participants were asked to analyze both maps manually designed by experienced cartographers and mapsautomatically generalized with the use of selected machine learning and deep learning models, namely random forest (RF), deep learning (DL), decision trees (DT) and decision trees optimized with genetic algorithms (DTGA). Based on four tasks and two datasets containing: source settlements, manually (human) and automatically generalized ones to smaller scales the users had to identify important settlement patterns and judge if the result was machine or human design. The experiment was conducted with the use of a dedicated web application. Additionally, eye-tracking data were recorded using a Tobii X2-30 eye-tracker. The preliminary results, as shown in Figure 1, suggest that the generalization results that successfully keep the specific settlement patterns are: 1) the automated results (AI generalization) with the use of random forest (RF) and deep learning (DL), and 2) the reference atlas map, designed by experienced cartographers (human generalization). In this preliminary study, participants found the decision tree (DT) results the least successful for maintaining the specific settlement patterns.

Additive manufacturing encompasses technologies that produce three-dimensional computer-aided design (CAD) models through a layer-by-layer production process. Compared to traditional manufacturing methods, additive manufacturing technologies offer significant advantages in producing intricate components with minimal energy consumption, reduced raw material waste, and shortened production timelines. AM methods based on shielded gas welding have recently piqued the interest of researchers due to their high efficiency and cost-effectiveness in manufacturing critical components. However, one of the most formidable challenges in additive manufacturing methods based on shielded gas welding lies in the irregularity of weld bead height at different points, compromising the precision of components produced using these techniques. In this current research, we aimed to achieve uniform weld heights along the welding path by considering the most influential parameters on weld bead geometry and conducting experimental tests. Input parameters of the process, including nozzle angle, welding speed, wire speed, and voltage, were considered. Simultaneously, image processing and wavelet transform were employed to assess the uniformity of weld bead height. These parameters were applied to produce intricate parts after identifying optimal parameters that yielded the smoothest weld lines. According to the results, the appropriate bead for manufacturing the part was extracted. The results show that the smoothest bead line is achieved in 27 V as the highest level of voltage, at a 90° nozzle position and the maximum wire feed rate. Parts manufactured using this method across different layers exhibited no distortions, and the repeatability of production substantiated the high reliability of this approach for component manufacturing. © 2024 by the authors.

In the rapidly evolving virtual reality (VR) field, assessing video quality on head-mounted displays (HMDs) for 360° videos presents unique opportunities and challenges. As immersive multimedia becomes increasingly widespread, understanding how users perceive and evaluate the quality of 360° videos is essential. This thesis investigates the subjective quality assessment tests of 360° videos, examining how participants' VR experiences, viewing conditions, and exploration behaviors shape perceptions of quality. This thesis aims to perform subjective quality assessment tests to study and understand how participants perceive the quality of 360° videos on an HMD. The thesis starts with an extended summary of the field of subjective quality assessment for 360° videos, followed by eight key publications, and unfolds into three main parts.

The first part of the thesis focuses on data collection to establish ground truth. It includes a comprehensive survey of annotated 360° images and videos datasets related to subjective quality assessment. It also presents a set of datasets collected specifically for subjective quality assessment tests for 360° videos with different test methods and viewing conditions conducted as part of the research. The second part of the thesis investigates how varying levels of VR experience affect participants' video quality assessments. It compares two test methods, the absolute category rating (ACR) and the modified ACR (MACR) method, to evaluate 360° video quality. Furthermore, this part evaluates simulator sickness in participants viewing 360° video on an HMD and explores how their prior VR experience levels correlate with the occurrence of these symptoms. The third and final part of the thesis focuses on assessing viewing conditions and rating consistency. It involves conducting subjective quality assessment tests for 360° videos under different viewing conditions, such as standing and seated viewing, and providing a statistical analysis of the psychophysical and psychophysiological measures. This part also investigates the consistency of 360° video quality assessments through repeated subjective quality assessment tests under opportunity-limited conditions. It examines how quality assessments vary between the standing and seated viewing conditions and explores whether participants' subjective evaluations of 360° videos change over time or remain stable across repeated exposures.

This paper presents a crucial multispectral image registration and sensor calibration method for an agricultural application. The multispectral images are obtained using a special drone equipped with multiple cameras flying at low altitudes. However, the distance between lenses, the lens distortions and the low-altitude flights lead to a lack of alignment in the built-in normalized difference vegetation index (NDVI). This lack of alignment results in a very poor performance in further analysis, especially for image segmentation and target detection to distinguish crops from invasive plants. In this work, we point out the importance of reducing this misalignment. To do so, the near-infrared and red sensors are first calibrated to remove the lens distortions. Then, the corresponding keypoints are utilized to calculate the transformation matrix and to minimize the back-projection error. The registered near-infrared and red images are then used to compute NDVI. The experimental results show higher alignment and F1-score of 0.73 which is a significant improvement in the performance of a trained deep neural network using NDVI in the detection of invasive plants. This is particularly a challenging task as the invasive plants resemble the desired crops. © 2024 IEEE.

Context: The research volume is continuously increasing. Manual analysis of large topic scopes and continuously updating literature studies with the newest research results is effort intensive and, therefore, difficult to achieve.

Objective: To discuss possibilities and next steps for using LLMs (e.g., GPT-4) in the mapping study process.

Method: The research can be classified as a solution proposal. The solution was iteratively designed and discussed among the authors based on their experience with LLMs and literature reviews.

Results: We propose strategies for the mapping process, outlining the use of agents and prompting strategies for each step.

Conclusion: Given the potential of LLMs in literature studies, we should work on a holistic solutions for LLM-supported mapping studies. 

The aim of the current study is to estimate whether spatial data on place features from OpenStreetMap (OSM) produce results similar to those when employing register data to predict future violent crime in public across three Swedish cities of varying sizes. Using violent crime in public as an outcome, four models for each city are produced using a Risk Terrain Modeling approach. One using spatial data on place features from register data and one from OSM, one model with prior violent crime excluded and one with prior crime included. The results show that several place features are significantly associated with violent crime in public independent of using register or OSM data as input. While models using register data seem to produce more accurate and efficient predictions than OSM data for the two smaller cities, the difference for the largest city is negligible indicating that the models provide similar results. As such, OSM place feature data may be of value when predicting the spatial distribution of future violent crime in public and provide results similar to those when using register data, at least when employed in larger compared to smaller cities. Possibilities, limitations, and avenues for future research when using OSM data in place-based criminological research are discussed. © The Author(s) 2024.

Mobile gaming accounts for more than 50% of global online gaming revenue, surpassing console and browser-based gaming. The success of mobile gaming titles depends on optimizing applications for the specific hardware constraints of mobile devices, such as smaller displays and lower computational power, to maximize battery life. Additionally, these applications must dynamically adapt to the variations in network speed inherent in mobile environments. Ultimately, user engagement and satisfaction are critical, necessitating a favorable comparison to browser and console-based gaming experiences. While Quality of Experience (QoE) subjective evaluations through user surveys are the most reliable method for assessing user perception, various factors, termed influence factors (IFs), can affect user ratings of stimulus quality. This study examines human influence factors in mobile gaming, specifically analyzing the impact of user delight towards displayed content and the effect of gaze tracking. Using Pupil Core eye-tracking hardware, we captured user interactions with mobile devices and measured visual attention. Video stimuli from eight popular games were selected, with resolutions of 720p and 1080p and frame rates of 30 and 60 fps. Our results indicate a statistically significant impact of user delight on the MOS for most video stimuli across all games. Additionally, a trend favoring higher frame rates over screen resolution emerged in user ratings. These findings underscore the significance of optimizing mobile gaming experiences by incorporating models that estimate human influence factors to enhance user satisfaction and engagement.

This letter introduces a novel regression model structure for the inflated Rayleigh distribution, which effectively models high dynamic amplitude pixel values in synthetic aperture radar (SAR) images. The proposed model estimates the mean of inflated Rayleigh distribution signals by a structure that includes a set of regressors and a link function. The inflated Rayleigh distribution combines the Rayleigh and a degenerate distribution, assigning non-null probability specifically for observed values equal to zero. Null pixel values in amplitude SAR images can be randomly distributed within the image, especially in low-intensity areas; a model capable of incorporating these values is essential to avoid changes in image statistics. Extensive evaluations are conducted using simulated and real SAR images to validate the proposed model, specifically focusing on ground-type detection in high dynamic amplitude pixel values scenarios. The performance of the proposed inflated Rayleigh regression model is compared with traditional Gaussian-based regression models, excelling in terms of ground-type detection in a SAR image obtained from the ICEYE radar. 

Background: 

The prevalence of Agile Software Development (ASD) practices has increased the prominence of individual and interpersonal skills. The human-centric nature of ASD practices makes it imperative for identifying and assigning capable professionals to constitute a workable team. Despite evidence from previous research in relation to the influence of workforce skillset and the effects of team climate factors on the performance of a team, the areas of capability measurement and factors influencing team climate in ASD remain largely unexplored. 

Objectives:

This thesis aims to aggregate evidence, from both former literature and current day practice, towards investigating capability measurement in ASD. Further, to address the gap in relation to team climate research in the ASD context, this thesis also investigates the effects of capability measures on team climate factors within industrial contexts.

Method:

A mixed-method approach was employed to address the thesis’ objectives, where a Systematic Literature Review (SLR) and multiple industrial surveys were conducted. A state of the practice survey (S1) was conducted to identify and gather evidence regarding capability measures relevant to the ASD context. To determine the relationship between capability measures pertaining to professionals and an agile team’s climate, first an industrial survey (S2) was carried out to study the influence of personality traits. Then, survey S2 was replicated twice (S3) and was further extended to examine the impact of other capability measures, in addition to personality traits (S4). 

Results:

Our SLR retrieved individual and team capability measures, where measures in relation to communication, interpersonal, and personal aspects were majorly emphasized by previous studies. Results from survey S1, where agile practitioners from multiple organizations participated, aligned with our SLR findings and practitioners perceived the majority of the SLR’s measures as relevant to ASD context. Further, the surveys conducted within a large-sized telecom company – S2, S3 and S4, brought to light multiple significant relationships that some of the capability measures showed with team climate factors. The meta-analytic effects observed by analyzing three samples gathered from surveys S2 and S3 showed that a rise in the neuroticism level of a person corresponded to a slight decline in the person’s perceived level of team climate. Further, our investigations identified that the inclusion of a wide range of capability measures, i.e., measures comprising both personality traits and other social aspects of capability measures, as input to regression models could explain more variance in the team climate factors.

Conclusion:

The empirical evidence gathered by employing mixed-methods and examining diverse organizational contexts, contributed towards better realization of capability measurement in ASD and identifying factors affecting agile team climate. The comprehensive list of capability measures acquired by our SLR were validated, through an industrial survey, by experienced agile practitioners who were associated with diverse roles and domains. This makes our SLR findings applicable to a wider audience. The findings from multiple surveys executed in industrial agile contexts showed that capability measures of team members contributed to a small, yet significant, portion of the variance in team climate factors, indicating the need to consider human factors and their effect upon team climate, and the need for gathering further data from diverse contexts, and perhaps to also include additional human factors. However, while applying the uncovered relationships to practice, one needs to evaluate whether they are valid (and to what degree). We believe a long-term inspection of capability measures can aid towards acquiring more data that would be necessary to establish robust team climate prediction models.

The concept of trust has been extensively explored by governments, researchers, and academic communities focusing on public authorities and the financial system, albeit in separate contexts. Trust plays a vital role in both sectors, influencing various aspects of governance, economic stability, and societal well-being. However, the relationship and interdependencies between trust in the government and trust in the financial system remain relatively unexplored. In addressing this gap, this study aims to improve the understanding of the role of trust in the socio-economic system and provide a framework for analysing the complex causal mechanisms between developments in the financial and public sectors using trust concepts. To achieve this, the study adopts the Fuzzy Cognitive Mapping (FCM) method in combination with the fuzzy Delphi method (FDM) as the methodological approach. The results highlight that even a small decline in trust can have severe repercussions on the stability of the financial system, deposit levels, exchange rate stability, and the prevalence of non-performing loans. Additionally, violations of trust in the financial sector also impact the development of the public sector, resulting in decreased trust in the government, fiscal stability, tax revenues, and government bond purchases. The study also demonstrated that when trust in both the financial sector and the government is eroded simultaneously, the complexities and the extent of negative consequences are amplified. These findings emphasize the interconnected nature of trust dynamics in both sectors and underscore the importance of a comprehensive approach to addressing trust-related challenges.