Search

Your search keyword '"Bosnar, Lovro"' showing total 13 results
Start Over Author "Bosnar, Lovro"
13 results on '"Bosnar, Lovro"'

1. Simulation of microstructures and machine learning

Author

Schladitz, Katja, Redenbach, Claudia, Barisin, Tin, Jung, Christian, Jeziorski, Natascha, Bosnar, Lovro, Fulir, Juraj, and Gospodnetić, Petra

Subjects
Computer Science - Computer Vision and Pattern Recognition, 60D05
Abstract

Machine learning offers attractive solutions to challenging image processing tasks. Tedious development and parametrization of algorithmic solutions can be replaced by training a convolutional neural network or a random forest with a high potential to generalize. However, machine learning methods rely on huge amounts of representative image data along with a ground truth, usually obtained by manual annotation. Thus, limited availability of training data is a critical bottleneck. We discuss two use cases: optical quality control in industrial production and segmenting crack structures in 3D images of concrete. For optical quality control, all defect types have to be trained but are typically not evenly represented in the training data. Additionally, manual annotation is costly and often inconsistent. It is nearly impossible in the second case: segmentation of crack systems in 3D images of concrete. Synthetic images, generated based on realizations of stochastic geometry models, offer an elegant way out. A wide variety of structure types can be generated. The within structure variation is naturally captured by the stochastic nature of the models and the ground truth is for free. Many new questions arise. In particular, which characteristics of the real image data have to be met to which degree of fidelity., Comment: Preprint of: K. Schladitz, C. Redenbach, T. Barisin, C. Jung, N. Jeziorski, L. Bosnar, J. Fulir, P. Gospodneti\'c: Simulation of Microstructures and Machine Learning, published in Continuum Models and Discrete Systems by F. Willot, J. Dirrenberger, S. Forest, D. Jeulin, A.V. Cherkaev (eds), 2024, Springer Cham. The final version is https://doi.org/10.1007/978-3-031-58665-1

Published
2025
Full Text
View/download PDF

2. SYNOSIS: Image synthesis pipeline for machine vision in metal surface inspection

Author

Fulir, Juraj, Jeziorski, Natascha, Bosnar, Lovro, Hagen, Hans, Redenbach, Claudia, Gospodnetić, Petra, Herrfurth, Tobias, Trost, Marcus, and Gischkat, Thomas

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Computational Engineering, Finance, and Science, Computer Science - Graphics, I.2.1, I.2.10, I.4.6, I.4.9, I.4.7, I.3.8, I.3.6, I.3.5, I.3.7, I.5.4, J.6, J.7
Abstract

The use of machine learning (ML) methods for development of robust and flexible visual inspection system has shown promising. However their performance is highly dependent on the amount and diversity of training data. This is often restricted not only due to costs but also due to a wide variety of defects and product surfaces which occur with varying frequency. As such, one can not guarantee that the acquired dataset contains enough defect and product surface occurrences which are needed to develop a robust model. Using parametric synthetic dataset generation, it is possible to avoid these issues. In this work, we introduce a complete pipeline which describes in detail how to approach image synthesis for surface inspection - from first acquisition, to texture and defect modeling, data generation, comparison to real data and finally use of the synthetic data to train a defect segmentation model. The pipeline is in detail evaluated for milled and sandblasted aluminum surfaces. In addition to providing an in-depth view into each step, discussion of chosen methods, and presentation of ML results, we provide a comprehensive dual dataset containing both real and synthetic images., Comment: Initial preprint, 21 pages, 21 figures, 6 tables

Published
2024

3. Simulation of Microstructures and Machine Learning

Author

Schladitz, Katja, Redenbach, Claudia, Barisin, Tin, Jung, Christian, Jeziorski, Natascha, Bosnar, Lovro, Fulir, Juraj, Gospodnetić, Petra, Willot, François, editor, Dirrenberger, Justin, editor, Forest, Samuel, editor, Jeulin, Dominique, editor, and Cherkaev, Andrej V., editor

Published
2024
Full Text
View/download PDF

7. Texture Synthesis for Surface Inspection

Author

Bosnar, Lovro, Rauhut, Markus, Hagen, Hans, Gospodnetic, Petra, Bosnar, Lovro, Rauhut, Markus, Hagen, Hans, and Gospodnetic, Petra

Abstract

The automated visual surface inspection planning is an important part of the quality assurance in automated custom product manufacturing. Visual surface inspection planning tackles image acquisition design and defect detection. Both tasks greatly benefit from the utilization of realistic and automated image synthesis of the inspected object. The realism of synthesized images greatly depends on object material, whose properties are largely influenced by texture. In this work, we focus on parametric texture synthesis and its application for visual surface inspection planning. We start by analyzing texture present on physical samples and introduce the requirements for texture synthesis models in visual surface inspection. Based on observation and surface characterization standards we present a model capable of reproducing texture on physical samples. This approach is generalized and further models are presented with respect to requirements. Finally, we highlight the importance of surface texture from the visual inspection planning perspective.

Published
2022

10. Modeliranje prostornog razvoja bioloških sustava

Author

Bosnar, Lovro and Domazet-Lošo, Mirjana

Subjects
Diffusion-limited aggregation model, Biofilm, TEHNIČKE ZNANOSTI. Računarstvo, model difuzno-ograničene agregacije, model kolonizacije prostora, Eden model, Space colonization model, modelling, modeliranje, Perlinov šum, TECHNICAL SCIENCES. Computing, vizualizacija, prostorni uzorci, Perlin noise, spatial patterns, visualization, Biofilm, Eden model, Diffusion-limited aggregation model, Space colonization model, Perlin noise, spatial patterns, modelling, visualization
Abstract

Cilj ovog rada bio je modelirati i vizualizirati prostorne uzorke određenog biološkog sustava. Odlučili smo modelirati i vizualizirati razvoj biofilma. Razvoj biofilma sadrži zanimljive oblike poput kanalića i nepravilne konture. Nadalje, proučili smo postojeće algoritme koji rezultiraju prostornim uzorcima. Odlučili smo se za korištenje Eden modela, modela difuzno-ograničene agregacije (Diffusion-limited aggregation), modela kolonizacije prostora (Space colonization algorithm) i Perlinovog šuma. Navedeni modeli predstavljaju osnovne građevne blokove kojim gradimo željeni model. Nadalje, implementirali smo navedene modele, eksperimentirali njima, te smo predstavili rezultate eksperimenata. Konačno, eksperimentirali smo s kombinacijama navedenih modela te smo predstavili rezultate. Najbolji model bio je odabran i opisan u detalje. The goal of this paper was to model and to visualize the spatial patterns of a certain biological system. We have decided that the biological system that we will use is the biofilm. The biofilm development contains interesting shapes such as channels and irregular contour. Next, we have studied the existing algorithms that produce spatial patterns. We have decided on using the Eden model, the Diffusion-limited aggregation model, the Space colonization model, and the Perlin noise. Those models were used as the basic building blocks for the desired model. Next, we have implemented those models, experimented and presented results. Finally, we have experimented with the combination of these models and we have presented the results. The best model was chosen and described in detail.

Published
2019

11. Analysis of Algorithms for Finding Maximum Flow in a Graph

Author

Bosnar, Lovro and Burić, Tomislav

Subjects
Ford-Fulkersonov algoritam, pretraživanje u dubinu, tok, TEHNIČKE ZNANOSTI. Računarstvo, graph, Ford-Fulkerson algorithm, Dijkstrin algoritam, Dijkstra's algorithm, TECHNICAL SCIENCES. Computing, flow, network, mreža, graf, breadth first search, depth first search, pretraživanje u širinu
Abstract

Cilj ovog rada bio je istražiti algoritam traženja maksimalnog toka u grafu, ali s različitim algoritmima traženja puta. Motivacija je bila potpunije upoznavanje maksimalnog toka i algoritama pretraživanja. U prvom dijelu je dana temeljna teorija grafova, objašnjen je specifičan graf – mreža te tok u mreži. U drugom dijelu je objašnjen algoritam za traženje maksimalnog toka. Specifično je obrađen Ford-Fulkersonov algoritam. Konačno u trećem dijelu prikazana je programska implementacija. Objašnjen je način prikaza grafa u računalu. Nakon svih potrebnih definicija analizirana su tri algoritma za traženje puta. Algoritam pretraživanja u širinu, algoritam pretraživanja u dubinu i Dijkstrin algoritam. Algoritmi su opisani matematički, pseudokodom te u dodatku se nalazi programska implementacija. Za svaki od algoritama navedene su iteracije i neke ključne karakteristike. Purpose of this paper was to study maximum flow algorithm with different path finding algorithms. Motivation was to learn more about maximum flow and path finding algorithms. The first part of this paper contains fundamentals of graph theory, theory behind specific graphs called networks and flow in network. The second part defines the core of this paper – Ford-Fulkerson maximum flow algorithm. Algorithm was described mathematically and with pseudo-code. The third part contains specifics of program implementation and analysis of maximum flow algorithm with different path finding algorithms. It was defined how to represent a graph in a computer. Chosen algorithms are: breadth first search, depth first search and Dijkstra's algorithm. Every path finding algoritam was described and for every algorithm the corresponding iterations were showed.

Published
2017

12. COMPARISON OF MAGENTA AND OPENCART ECOMMERCE PLATFORM

Author

Bosnar, Lovro and Kežmah, Boštjan

Subjects
izgradnja spletne trgovine, brezplačna platforma, online store, building an online store, udc:004.777(043.2), OpenCart, spletna trgovina, free of charge platform, Magento
Abstract

V sodobnem času veliko ljudi nakupuje preko spleta, zato je hitrost delovanja in razvoj spletnih trgovin zelo pomembna. Razvijalci spletne trgovine večinoma ne razvijajo samostojno oziroma od začetka, temveč posežejo po vnaprej pripravljenih platformah za izgradnjo spletne trgovine in jo razvijejo ter preoblikujejo po željah lastnika oziroma naročnika. V diplomskem delu bomo izdelali dve spletni trgovini v platformah Magento in OpenCart. Primerjali bomo hitrost in težavnost razvoja in nastavitev osnovnih funkcij za izgradnjo spletne trgovine. In modern times many people shop online, so the spped of service and development of online stores are of great importance. Developers usually do not develop an independent online store or start building one from the beginning, but resort to ready – made platforms for building online stores. The store is then developed and transformed according to the wishes of the owner or client. In this thesis we will develop two online shops in the platforms Magento in OpenCart. We will compare the speed and difficulty of developing and setting up the basic functions of building an online store.

Published
2015

13. Procedural Modeling and Image Synthesis for Virtual Surface Inspection Planning

Author

Bosnar, Lovro and Bosnar, Lovro

Abstract

Product manufacturing is performed in a massively automated and increasingly customized manner. However, overall production speed is limited by automation of inspection since each product has to ensure the required quality. A widespread and often-used quality assurance method is visual surface inspection. Automated surface inspection relies on an inspection plan and defect recognition algorithms. Both inspection planning and defect recognition algorithms development heavily rely on the availability of representative image data containing various product surface textures and imperfections showing a wide variety of possible surface responses to different viewing and lighting conditions. Due to the advancements in manufacturing, defects in products occur rarely, with different frequencies of appearance, followed by a subjective and laborious annotation process. Further, since the surface texture is often not relevant to product performance and thus not controlled, products with different surface textures are not treated as different product samples and thus not provided. Motivated by aforementioned problems, this work introduces the following contributions: (1) image synthesis requirements for industrial quality inspection and a novel realistic image synthesis pipeline satisfying those requirements (Chapter 4), (2) texture synthesis requirements for industrial quality inspection and a procedural approach to parameterized surface texture modeling incorporating domain knowledge (Chapter 5) and (3) defect synthesis requirements for industrial quality inspection as well as a procedural approach to parameterized defect modeling (Chapter 6). The contributions presented in this thesis, make it possible to obtain, in a controllable and automated manner, the required amount of image data, containing realistic and varying surface textures resembling machining surfaces as well as diversified geometrical defects with automated, pixel-precise annotations (Chapters 7,8). The present

Catalog

Books, media, physical & digital resources
See catalog results