1. Background purification framework with extended morphological attribute profile for hyperspectral anomaly detectionJu Huang, Kang Liu, Mingliang Xu, Matjaž Perc, Xuelong Li, 2021, original scientific article Abstract: Hyperspectral anomaly detection has attracted extensive interests for its wide use in military and civilian fields, and three main categories of detection methods have been developed successively over past few decades, including statistical model-based, representation-based, and deep-learning-based methods. Most of these algorithms are essentially trying to construct proper background profiles, which describe the characteristics of background and then identify the pixels that do not conform to the profiles as anomalies. Apparently, the crucial issue is how to build an accurate background profile; however, the background profiles constructed by existing methods are not accurate enough. In this article, a novel and universal background purification framework with extended morphological attribute profiles is proposed. It explores the spatial characteristic of image and removes suspect anomaly pixels from the image to obtain a purified background. Moreover, three detectors with this framework covering different categories are also developed. The experiments implemented on four real hyperspectral images demonstrate that the background purification framework is effective, universal, and suitable. Furthermore, compared with other popular algorithms, the detectors with the framework perform well in terms of accuracy and efficiency.
Keywords: detectors, anomaly detection, image reconstruction, hyperspectral imaging, training, optics, dictionaries, background purification, extended attribute profile, sparse representation, stacked autoencoder
Published in DKUM: 19.08.2024; Views: 92; Downloads: 7
 Full text (5,36 MB)
This document has many files! More... |
2. DigiPig : First developments of an automated monitoring system for body, head and tail detection in intensive pig farmingMarko Ocepek, Anja Žnidar, Miha Lavrič, Dejan Škorjanc, Inger Lise Andersen, 2022, original scientific article Abstract: The goal of this study was to develop an automated monitoring system for the detection of pigs’ bodies, heads and tails. The aim in the first part of the study was to recognize individual pigs (in lying and standing positions) in groups and their body parts (head/ears, and tail) by using machine learning algorithms (feature pyramid network). In the second part of the study, the goal was to improve the detection of tail posture (tail straight and curled) during activity (standing/moving around) by the use of neural network analysis (YOLOv4). Our dataset (n = 583 images, 7579 pig posture) was annotated in Labelbox from 2D video recordings of groups (n = 12–15) of weaned pigs. The model recognized each individual pig’s body with a precision of 96% related to threshold intersection over union (IoU), whilst the precision for tails was 77% and for heads this was 66%, thereby already achieving human-level precision. The precision of pig detection in groups was the highest, while head and tail detection precision were lower. As the first study was relatively time-consuming, in the second part of the study, we performed a YOLOv4 neural network analysis using 30 annotated images of our dataset for detecting straight and curled tails. With this model, we were able to recognize tail postures with a high level of precision (90%).
Keywords: pig, welfare, image processing, object detection, deep learning, smart farming
Published in DKUM: 11.07.2024; Views: 87; Downloads: 3
Full text (48,11 MB)
This document has many files! More... |
3. Geometric Shape Characterisation Based on a Multi-Sweeping ParadigmBorut Žalik, Damjan Strnad, David Podgorelec, Ivana Kolingerová, Andrej Nerat, Niko Lukač, Štefan Kohek, Luka Lukač, 2023, original scientific article Keywords: computer science, image analysis, computational geometry, local reflection symmetry
Published in DKUM: 24.05.2024; Views: 300; Downloads: 11
Full text (1,87 MB)
This document has many files! More... |
4. Evaluating seagrass meadow dynamics by integrating field-based and remote sensing techniquesDanijel Ivajnšič, Martina Orlando-Bonaca, Daša Donša, Jaša Veno Grujić, Domen Trkov, Borut Mavrič, Lovrenc Lipej, 2022, original scientific article Abstract: Marine phanerogams are considered biological sentinels or indicators since any modification in seagrass meadow distribution and coverage signals negative changes in the marine environment. In recent decades, seagrass meadows have undergone global losses at accelerating rates, and almost one-third of their coverage has disappeared globally. This study focused on the dynamics of seagrass meadows in the northern Adriatic Sea, which is one of the most anthropogenically affected areas in the Mediterranean Sea. Seagrass distribution data and remote sensing products were utilized to identify the stable and dynamic parts of the seagrass ecosystem. Different seagrass species could not be distinguished with the Sentinel-2 (BOA) satellite image. However, results revealed a generally stable seagrass meadow (283.5 Ha) but, on the other hand, a stochastic behavior in seagrass meadow retraction (90.8 Ha) linked to local environmental processes associated with anthropogenic activities or climate change. If systemized, this proposed approach to monitoring seagrass meadow dynamics could be developed as a spatial decision support system for the entire Mediterranean basin. Such a tool could serve as a key element for decision makers in marine protected areas and would potentially support more effective conservation and management actions in these highly productive and important environments.
Keywords: Adriatic Sea, seagrass meadow, change analysis, Cimodocea nodosa, image classifiers, Sentinel-2, marine biology, hydrobiology
Published in DKUM: 17.05.2024; Views: 152; Downloads: 9
Full text (3,45 MB)
This document has many files! More... |
5. A VAN-Based Multi-Scale Cross-Attention Mechanism for Skin Lesion Segmentation NetworkShuang Liu, Zeng Zhuang, Yanfeng Zheng, Simon Kolmanič, 2023, original scientific article Abstract: With the rise of deep learning technology, the field of medical image segmentation has undergone rapid development. In recent years, convolutional neural networks (CNNs) have brought many achievements and become the consensus in medical image segmentation tasks. Although many neural networks based on U-shaped structures and methods, such as skip connections have achieved excellent results in medical image segmentation tasks, the properties of convolutional operations limit their ability to effectively learn local and global features. To address this problem, the Transformer from the field of natural language processing (NLP) was introduced to the image segmentation field. Various Transformer-based networks have shown significant performance advantages over mainstream neural networks in different visual tasks, demonstrating the huge potential of Transformers in the field of image segmentation. However, Transformers were originally designed for NLP and ignore the multidimensional nature of images. In the process of operation, they may destroy the 2D structure of the image and cannot effectively capture low-level features. Therefore, we propose a new multi-scale cross-attention method called M-VAN Unet, which is designed based on the Visual Attention Network (VAN) and can effectively learn local and global features. We propose two attention mechanisms, namely MSC-Attention and LKA-Cross-Attention, for capturing low-level features and promoting global information interaction. MSC-Attention is designed for multi-scale channel attention, while LKA-Cross-Attention is a cross-attention mechanism based on the large kernel attention (LKA). Extensive experiments show that our method outperforms current mainstream methods in evaluation metrics such as Dice coefficient and Hausdorff 95 coefficient.
Keywords: CNNs, deep learning, medical image processing, NLP, semantic segmentation
Published in DKUM: 14.03.2024; Views: 496; Downloads: 306
Full text (1,46 MB)
This document has many files! More... |
6. Literary tourism : the role of Russian 19th century travel literature in the positioning of the smallest European royal capital - CetinjeAndriela Vitić-Ćetković, Ivona Jovanović, Jasna Potočnik Topler, 2020, original scientific article Abstract: Increasing competition on the global tourism market forces numerous tourist destinations to create a specific image and diversify their offers through innovative and sustainable tourism products. In view of this, there is a range of possibilities for utilizing historical resources, tangible and intangible cultural heritage, including travel literature, which has not been considered in Montenegro thus far in the context of potentials for enhancing the tourist offer. This paper is focusing on the research of travel literature by 19th century authors from Russia who wrote about Cetinje and Montenegro, as well as the possibilities of creating a destination image and diversified experience for specific market niches, primarily the tourists from Russia who have been among the most numerous in Montenegro. The expected outcome of the research is to point out the importance of valorisation of Russian traveler literature in the context of creating a destination image. Considering the negative propaganda of a part of the media in Russia when Montenegro entered NATO in 2017, it is expected that this Balkan and Adriatic country, whose primary business is tourism, will have to identify and acquire state-of-the-art modalities for attracting new target segments from the Russian Federation. In this paper, literary tourism with the concepts of town - museum, town of books and storytelling, as marketing communications tools, are proposed to promote the revival and valorization of historical events, historical figures and Cetinje's former image in the positioning of the tourism destination, also in the conte xt of sustainable tourism development.
Keywords: travel literature, literary tourism, heritage, destination image, destination positioning, storytelling, Monte Negro, Cetinje
Published in DKUM: 26.01.2024; Views: 235; Downloads: 4
Full text (2,75 MB)
This document has many files! More... |
7. The influence of façade colour, glazing area and geometric configuration of urban canyon on the spectral characteristics of daylightNataša Šprah, Jaka Potočnik, Mitja Košir, 2024, original scientific article Abstract: Since the non-image-forming (NIF) effect of daylight on the human circadian system is widely accepted, adequate exposure to daylight is now considered one of the elements of a healthy life. In urban environment, one of the prerequisites for adequately lit indoor spaces is the amount and quality of daylight reaching the window, which is highly dependent on the characteristics of the urban environment. The aim of the study was to determine whether there are correlations between urban density (i.e., distance between buildings, building height), façade surface characteristics (i.e., colour and Window-to-Wall Ratios – WWR) and NIF potential of daylight. The study was conducted on a parametric geometric model of a street canyon covering a wide range of characteristics. Simulation results were interpreted using the Relative Melanopic Efficacy coefficient and Sky View Factor. The results indicate that the colour of the opposite façade can substantially impact the resulting NIF potential, especially for orange-red or blue hues. The results of the study show that this influence for building heights between 3 and 8 storeys becomes significant when the width of the urban canyon is less than 25 m and becomes substantially smaller at WWRs above 30 %.
Keywords: urban planning, daylight, non-image forming effects of light, circadian light, urban canyon, façade colour
Published in DKUM: 19.01.2024; Views: 415; Downloads: 21
Full text (14,66 MB)
This document has many files! More... |
8. High strain-rate deformation analysis of open-cell aluminium foamAnja Mauko, Mustafa Sarıkaya, Mustafa Güden, Isabel Duarte, Matej Borovinšek, Matej Vesenjak, Zoran Ren, 2023, original scientific article Abstract: This study investigated the high-strain rate mechanical properties of open-cell aluminium foam M-pore®. While previous research has examined the response of this type of foam under quasi-static and transitional dynamic loading conditions, there is a lack of knowledge about its behaviour under higher strain rates (transitional and shock loading regimes). To address this gap in understanding, cylindrical open-cell foam specimens were tested using a modified Direct Impact Hopkinson Bar (DIHB) apparatus over a wide range of strain rates, up to 93 m/s. The results showed a strong dependency of the foam's behaviour on the loading rate, with increased plateau stress and changes in deformation front formation and propagation at higher strain rates. The internal structure of the specimens was examined using X-ray micro-computed tomography (mCT). The mCT images were used to build simplified 3D numerical models of analysed aluminium foam specimens that were used in computational simulations of their behaviour under all experimentally tested loading regimes using LS-DYNA software. The overall agreement between the experimental and computational results was good enough to validate the built numerical models capable of correctly simulating the mechanical response of analysed aluminium foam at different loading rates.
Keywords: Open-cell aluminium foam, Micro-computed tomography, High-strain rate, Direct impact hopkinson bar, Digital image correlation, Computer simulation
Published in DKUM: 06.12.2023; Views: 428; Downloads: 35
Full text (3,28 MB)
This document has many files! More... |
9. Design of an Embedded Position Sensor with Sub-mm Accuracy : magistrsko deloMatej Nogić, 2019, master's thesis Abstract: This master’s thesis presents the development of a machine-vision based localization unit developed at Robert Bosch GmbH, Corporate Sector Research and Advance Engineering in Renningen, Germany. The localization unit was developed primarily for position detection purposes with three degrees of freedom in highly versatile manufacturing systems but has an immense potential to be used anywhere where a precise, low-cost localization method on a two-dimensional surface is required. The complete product development cycle was carried out, from the components selection, schematic and optical system design, to the development of machine vision algorithms, four-layer Printed Circuit Board design and evaluation using an industrial robot. Thanks to the use of a patented two-dimensional code pattern, the localization unit can cover a surface area of 49 km2. The size and speed optimized, self-developed machine-vision algorithms running on a Cortex-M7 microcontroller allow achieving an accuracy of 100 µm and 60 Hz refresh rate.
Keywords: localization, machine-vision, code pattern, image sensor, embedded system
Published in DKUM: 14.01.2020; Views: 1340; Downloads: 52
Full text (18,20 MB) |
10. Vergleich des Images der deutschen Sprache im dritten Bildungsabschnitt an zwei verschiedenen Grundschulen in PrekmurjeMatija Toth, 2019, undergraduate thesis Abstract: Im Mittelpunkt der Diplomarbeit mit dem Titel ‚‚Vergleich des Images der deutschen Sprache im dritten Bildungsabschnitt an zwei verschiedenen Grundschulen in Prekmurje‘‘ steht das Image der deutschen Sprache bzw. ihr Vergleich an zwei, von Hinsicht des mehrsprachigen Unterrichtes, abgrenzenden Grundschulen in Slowenien.
Das Ziel der Untersuchung ist der Vergleich des Images der deutschen Sprache zwischen zwei Befragungsgruppen, die jeweils Schüler der einsprachigen Grundschule in Murska Sobota oder der zweisprachigen Grundschule in Lendava sind und sich im Alter zwischen zwölf und fünfzehn Jahren befinden. Die Forschung fokussiert sich auf die eventuellen Unterschiede zwischen den Sprachimages der zwei untersuchten Gruppen.
Im theoretischen Teil wird definiert, was eigentlich unter dem Begriff Image verstanden wird, von welchen Faktoren das Image einer Sprache beeinflusst wird, was die verbreitetsten Mythen über Sprachen sind und welche Rolle die englische Sprache dabei spielt.
Im empirischen Teil wird eine Analyse des Fragebogens folgen, das dieses Thema behandelt. An der Umfrage nahmen 149 Schüler des dritten Bildungsabschnittes aus Lendava und 139 aus Murska Sobota teil. Insgesamt wurde der Fragebogen von 278 Schülern ausgefüllt. Der Fragebogen beinhaltet Aufgabenstellungen und Fragen, mit Hilfe dessen man sich ein Bild über das Image des Deutschen, bei den Schülern der jeweiligen Schulen, machen konnte. Von großer Bedeutung war es, die subjektiven Informationen der Individuen zu bekommen, um so zu einem authentischen Resultat der wissenschaftlichen Arbeit zu kommen.
Keywords: Image der deutschen Sprache, mehrsprachiger Unterricht, Vergleich zweier Schulen
Published in DKUM: 16.10.2019; Views: 1317; Downloads: 202
Full text (1,21 MB) |
