1. Optimizing digital image quality for improved skin cancer detectionBogdan Dugonik, Marjan Golob, Marko Marhl, Aleksandra Vučinič Dugonik, 2025, izvirni znanstveni članek Opis: The rising incidence of skin cancer, particularly melanoma, underscores the need for improved diagnostic tools in dermatology. Accurate imaging plays a crucial role in early detection, yet challenges related to color accuracy, image distortion, and resolution persist, leading to diagnostic errors. This study addresses these issues by evaluating color reproduction accuracy across various imaging devices and lighting conditions. Using a ColorChecker test chart, color deviations were measured through Euclidean distances (∆E*, ∆C*), and nonlinear color differences (∆E00, ∆C00), while the color rendering index (CRI) and television lighting consistency index (TLCI) were used to evaluate the influence of light sources on image accuracy. Significant color discrepancies were identified among mobile phones, DSLRs, and mirrorless cameras, with inadequate dermatoscope lighting systems contributing to further inaccuracies. We demonstrate practical applications, including manual camera adjustments, grayscale reference cards, post-processing techniques, and optimized lighting conditions, to improve color accuracy. This study provides applicable solutions for enhancing color accuracy in dermatological imaging, emphasizing the need for standardized calibration techniques and imaging protocols to improve diagnostic reliability, support AI-assisted skin cancer detection, and contribute to high-quality image databases for clinical and automated analysis. Ključne besede: dermoscopy, melanoma, color analysis, color error, spectral power distribution, grey card, digital imaging standards Objavljeno v DKUM: 08.04.2025; Ogledov: 0; Prenosov: 4
Celotno besedilo (5,63 MB) |
2. Sequence-to-Sequence models and their evaluation for spoken language normalization of SlovenianMirjam Sepesy Maučec, Darinka Verdonik, Gregor Donaj, 2024, izvirni znanstveni članek Opis: Sequence-to-sequence models have been applied to many challenging problems, including
those in text and speech technologies. Normalization is one of them. It refers to transforming
non-standard language forms into their standard counterparts. Non-standard language forms
come from different written and spoken sources. This paper deals with one such source, namely
speech from the less-resourced highly inflected Slovenian language. The paper explores speech
corpora recently collected in public and private environments. We analyze the efficiencies of three
sequence-to-sequence models for automatic normalization from literal transcriptions to standard
forms. Experiments were performed using words, subwords, and characters as basic units for
normalization. In the article, we demonstrate that the superiority of the approach is linked to the
choice of the basic modeling unit. Statistical models prefer words, while neural network-based
models prefer characters. The experimental results show that the best results are obtained with
neural architectures based on characters. Long short-term memory and transformer architectures
gave comparable results. We also present a novel analysis tool, which we use for in-depth error
analysis of results obtained by character-based models. This analysis showed that systems with
similar overall results can differ in the performance for different types of errors. Errors obtained with
the transformer architecture are easier to correct in the post-editing process. This is an important
insight, as creating speech corpora is a time-consuming and costly process. The analysis tool also
incorporates two statistical significance tests: approximate randomization and bootstrap resampling.
Both statistical tests confirm the improved results of neural network-based models compared to
statistical ones. Ključne besede: low-resource language, applications, spoken language, normalization, character unit, subword unit, statistical model, long short-term memory, transformer, error analysis Objavljeno v DKUM: 31.01.2025; Ogledov: 0; Prenosov: 12
Celotno besedilo (437,99 KB) Gradivo ima več datotek! Več... |
3. Analysis and improvements of the mechanisms for cross-border interchange and activation of the regulating reserves : doctoral dissertationMarcel Topler, 2022, doktorska disertacija Opis: This Doctoral Thesis deals with the mechanisms for cross-border interchange and activation of the regulating reserves (RRs), i.e., Imbalance Netting Process (INP) and Cross-Border Activation of the RRs (CBRR), between participating Control Areas (CAs), to reduce the costs of balancing energy.
The main objective of INP is to interchange the RRs between participating CAs with opposite signs of interchange power variation. In comparison, the main objective of CBRR is to activate the RRs in participating CAs with the same signs of interchange power variation. Both the INP and CBRR aim to release the RRs and reduce balancing energy as part of the power system's safe operation.
The Thesis's main objective is to analyze the impact of the mechanisms for cross-border interchange and activation of the RRs on mutual oscillations of participating CAs and stability for small disturbances. The Thesis's secondary objective is to analyze the impact of the INP and CBRR on frequency quality, on the provision of Load-Frequency Control (LFC), on balancing energy and unintended exchange of energies between participating CAs.
Frequency quality in Continental Europe (CE) has been declining in recent years, so it is important that the mechanisms for cross-border interchange and activation of the RRs do not further impair its quality.
Both the classic INP and CBRR include a frequency-dependent contribution and, therefore, inherently affect the frequency response of the participating CAs, which is not discussed in the literature. Thus, the impact of the classic INP and CBRR on frequency quality and the provision of LFC is thoroughly evaluated with dynamic simulations of a three-CA test system and eigenvalue analysis of a two-CA system. It is demonstrated that both the classic INP and CBRR reduce the damping of the entire power system.
Therefore, a modified implementation of the classic INP and CBRR is presented, and improved INP and CBRR are proposed, which have no impact on the mutual oscillations of participating CAs and stability for small disturbances.
Furthermore, the dynamic simulations results confirm that the frequency quality can be improved by the classic INP and CBRR, although there are also cases where it can deteriorate. However, the improved INP and CBRR generally improve the frequency quality in all cases. The improved INP and CBRR also enhance the provision of LFC compared to the classic INP and CBRR. Moreover, the improved INP and CBRR reduce the unintended exchange of energies, thus increasing the economic effects of the INP's and CBRR's activation.
The improved INP increases energy exchange, therefore positive economic benefits can be expected in comparison to the system with the classic INP.
However, the improved CBRR reduces energy exchange, therefore positive economic benefits can be expected in comparison to the system with the classic CBRR, since energy exchange is paid by CA via bidding process. Ključne besede: load-frequency control, imbalance netting, cross-border activation, balancing
energy, regulating reserves, eigenvalue analysis, performance indicators, area control error, rate of change of frequency, control area Objavljeno v DKUM: 09.03.2023; Ogledov: 594; Prenosov: 121
Celotno besedilo (14,64 MB) |
4. The Influence of Styrian and Upper Carniola Dialects on the Pronunciation of English of Slovene StudentsDominika Podvratnik, 2018, magistrsko delo Opis: The master’s thesis considers the influence of Styrian and Upper Carniola dialect on the pronunciation of English of Slovene students. Various factors influence the Slovene students’ pronunciation of English. The influence of mother tongue on the Slovene students’ pronunciation of English has been firmly established. However, it is only logical to think that dialect also has an important influence on the Slovene students’ pronunciation of English. The purpose of the research is to determine the influence of Styrian dialect and Upper Carniola dialect on the pronunciation of English of Slovene students. The master’s thesis strengthens the evidence for dialect influence, observes the extent of dialect influence in both dialect areas, determines in which area the dialect influence on the pronunciation of English is more visible and identifies the reasons for such influence. The research employs various research methods. Contrastive analysis is used to compare the sound systems of Upper Carniola and Styrian dialect with the sound system of British English. Based on the findings of the contrastive analysis words are chosen and organised in a test which is used for obtaining samples for analysis. Fieldwork is then carried out. Auditory analysis of samples includes listening to audio material and producing phonemic transcriptions, which concludes with statistical analysis of data. The results show that both Upper Carniola and Styrian dialect significantly influence the pronunciation of English of Slovene students. Dialect influence is slightly more prevalent in the Styrian dialect region, especially in the case of RP English /e/ and /ʊ/. Slovene students of English speak interlanguage and their interphonology rules mostly stem from English phonological rules. Ključne besede: Upper Carniola dialect, Styrian dialect, contrastive analysis, error analysis, interlanguage Objavljeno v DKUM: 23.10.2018; Ogledov: 1431; Prenosov: 167
Celotno besedilo (1,06 MB) |
5. Foreign students' speech in EnglandManca Potušek, 2016, diplomsko delo Opis: The diploma thesis focuses on the influence of the French, Italian and Spanish language on the Standard Southern British English pronunciation. The research is based on the recordings of 10 French, 9 Italian and 9 Spanish students who spent half a year in England as Erasmus exchange students.
The theoretical part starts with an introductory chapter describing a short history of all four languages, as well as the position of the English language in the world and in the European Union. East Midlands English and Standard Southern British English are then presented with their phonological characteristics. Furthermore, the vowel systems of French, Italian and Spanish language are presented together with the most common errors the speakers of these languages encounter when pronouncing English vowels. In the following chapter the term “interlanguage” and research methods as Contrastive Analysis and Error Analysis are described. In the last segment of the theoretical part we are dealing with so-called formants which are frequencies that the vowels are made of and are essential for identification of a vowel.
In the empirical part, Error Analysis is used as a research method. The key part of the research is the analysis of the recordings of foreign students’ pronunciation of 11 SSBE vowels. The results of the research clearly prove a strong influence of the French, Italian, and Spanish accents on the pronunciation of SSBE vowels, as well as a great impact of General American. To analyse the data, the descriptive method was used.
The findings of the dissertation could be of help to other researchers or English teachers should they find themselves teaching in France, Italy or Spain. Ključne besede: Standard Southern British English, vowels, formants, foreign students, Error Analysis Objavljeno v DKUM: 08.09.2016; Ogledov: 1601; Prenosov: 92
Celotno besedilo (1,98 MB) |