Image classification systems have been widely applied in smartphones, industrial inspection, and automated devices. However, in real-world scenarios, existing classifiers often produce inconsistent or incorrect predictions due to factors such as image degradation, inter-class similarity, and model instability, thereby reducing the overall performance of the system. To address these issues and enha

Detta examensarbete syftar till att utvärdera om användningen av historisk testdata kan bidra till en förbättrad testprocess. I rapporten presenteras två JavaScript-algoritmer utvecklade av uppsatts- författarna, createTestRun och updateTestRun, som är specifikt utvecklade för att optimera urvalet av testfall och minimera redundans mellan testkörningar. Algoritmernas bygger på historisk data som hThis thesis presents an approach to utilize historical test execution data for improving the test process within a Quality Assurance (QA) department. The thesis research introduces two custom-built JavaScript algorithms, createTestRun and updateTestRun, designed to optimize the selection of test cases and effectively reduce redundancy in between test runs. The algorithms were developed by the thes

Denna kandidatuppsats undersöker hur användare uppfattar så kallade ”deceptive patterns” - vilseledande designmönster - i digitala användargränssnitt, samt var gränsen går mellan påverkan och manipulation. Studien kombinerar en kvantitativ enkätundersökning med en kvalitativ intervjustudie för att fånga både bredd och djup i ämnet. Enkäten, som riktades till svenska internetanvändare, inkluderade

Syftet med denna studie är att skapa en djupare förståelse och klargöra hur generativ AI kan appliceras som ett hjälpmedel för entreprenörer i anbudsprocessen. Offentliga upphandlingar präglas ofta av omfattande informationsmängder, komplexa formella krav och strikta tidsramar. När förfrågningsunderlagen dessutom är bristfälliga eller otydliga blir varje anbud en resurskrävande investering som måsThe purpose of this study is to develop a deeper understanding and clarify how generative AI can be applied as a support tool for contractors in the bidding process. Public procurements are often characterized by extensive amounts of information, complex formal requirements, and strict timeframes. When the tender documents are additionally incomplete or unclear, each bid becomes a resource-intensi

The Integration of Machine Learning in OFDM Receivers The integration of Machine Learning (ML) into various fields has prompted exploration into the effectiveness of advanced ML architectures within Orthogonal Frequency Division Multiplexing (OFDM) receivers. Prior work by Nvidia demonstrated the promise of Graph Neural Networks (GNNs) in replacing key components of a traditional receiver, surpas

Accurate modeling of wireless channels in urban environments is crucial for the design and evaluation of communication systems. Geometry-Based Stochastic Channel Models (GSCMs) provide a balance between physical retings—an oversimplification that can reduce model fidelity.alism and computational efficiency by simulating Multi-Path Components (MPCs) through simplified ray tracing and stochastic sca

There is a lack of research on how young adults experienced having a parent with a dementia-related illness, particularly regarding the support they received. This knowledge gap risked leaving young relatives without adequate help, as existing support systems were often designed for older relatives. Therefore, it was essential to explore their needs and experiences. The aim of this study was to ex

This study aimed to explore how individuals with experience of homelessness perceived the available support services, as well as how they experienced their encounters with social workers, highlighting effective current practices and identifying areas needing improvement. Based on narrative interviews with six formerly homeless individuals in Stockholm, Sweden, their experiences were analysed using

Globally, automated ecological cameras (PhenoCam) are widely used to monitor vegetation and seasonal changes. Snow is usually easy to identify, but image quality flags (like haze, glare) can make it hard to detect in large datasets. Manual quality control becomes prohibitively time-intensive for large-scale phenological studies, creating a critical need for robust automated snow detection methods.

Warming due to global climate change has led to various natural anomalies, such as increases in forest fires, water level rise and droughts. A region being affected heavily by climate change is the Colorado River Basin, having an arid, semi-arid climate with the Colorado Mountain snowmelt supplying Colorado river, which flows throughout the basin, accumulating water in Lake Mead Reservoir. The Bas

Approximately 80% of agricultural land in the world is comprised of annual crops which have a life span of only one year. Research has shown that these crops are unsustainable because of these short life spans. Soil erosion, nutrient leaching, and losses of soil carbon are just a few problems caused by annual monocultures. Additionally, with the changing climate, these cropping methods are both vu

Equity Real Estate Investment Trusts (REITs) face persistent challenges in performance tracking and financial reporting due to fragmented data systems, manual processes, and delayed access to critical insights. This thesis examines the use of Power BI, a business intelligence (BI) tool, as a means to modernize reporting practices and support data-driven decision-making within REITs. Drawing on use

In ongoing research, several concepts of “next-generation” cross laminated timber (CLT) are being investigated aiming at creating innovative and optimized CLT-alike products. In the project InnoTLT (“Innovative Tailored LaminatedTimber”) one main aim is to show how today’s CLT production could be made more material efficient by introducing alternative geometries, including large gaps between lamin

The legal opacity of AI technologies has long posed challenges in addressing algorithmic harms, as secrecy enables companies to retain competitive advantages while limiting public scrutiny. In response, ideas such as qualified transparency have been proposed to provide AI accountability within the confidentiality constraints. With the introduction of the EU AI Act, the foundations for human-centri

Railway maintenance is crucial for ensuring system availability, punctuality, safety, and comfort. Though studies have shown that in reality, maintenance is 'imperfect'-inspections do not reveal all defects and repairs do not return components to as good as new and may even leave them in a worse state, railway maintenance planning is still based on periodic block-type and age-based maintenance pol

Vehicle models are often bound by non-disclosure agreements with manufacturers.This thesis presents a vehicle model based on vehicle data gathered for the NationalAdvanced Driving Simulator in Ohio and therefore it is not bound by such limitationsand can be used in future research projects.The development and validation of the vehicle model requires understanding of thesubsystems studied, componen

Optimization of machining processes, such as milling, is essential for industrial efficiency and product quality. To achieve greater efficiency, it is necessary to understand how tools wear down in different conditions in order to anticipate possible undesirable events like sudden breakage or unpredictable degradation. This study focuses on understanding tool wear in dry milling of compacted graph

The European Union’s aim to decarbonize the building stock by 2050 requires significant reductions in the energy use for space heating and cooling. One such solution to decarbonization could be energy sharing between buildings. This paper presents the potential for energy saving and subsequent reduction in carbon emission with a low temperature district network in an energy community based in Malm