|
|
Creation and Demonstration of Assets for VR Application
Zouhar, Marek ; Tóth, Michal (referee) ; Milet, Tomáš (advisor)
This thesis deals with the concept of virtual reality, its history, present-day possibilities and available devices and technologies for virtual reality and ways of creating assets such as models, textures and animations for virtual reality applications. The practical part of this work deals with design and creation of three-dimensional models, textures, animations and environment for use in interactive application in virtual reality and also with design and creation of such application to demonstrate their use.
|
|
|
Character Modeling - Polygonal Wrapper
Žák, Pavel ; Herout, Adam (referee) ; Chudý, Robert (advisor)
This project is engaged in optimalization of 3D polygonal models. Main automatic and also manual principles and methods used in the area of character model optimalization are introduced and discussed. Next the approach named geometry mapping, which was created as a part of the project and enables the creation of models with desired topology, is described.
|
|
|
Deep Neural Networks for Landmark Detection on 3D Models
Kubík, Tibor ; Kodym, Oldřich (referee) ; Španěl, Michal (advisor)
Detekcia významných bodov je častým krokom pri analýze medicínskych dát. Čoraz bežnejšie sú tieto dáta reprezentované vo forme 3D modelov, príkladom sú povrchové skeny zubného oblúka pacienta. Hlboké neurónové siete sú vhodný spôsob, ako detekovať významné body v obraze. V prípade 3D dát je však toto spracovanie časovo i pamäťovo náročné, čo nevyhovuje požiadavkám kladeným medicínskymi aplikáciami. V tejto práci navrhujem metódu, ktorá tento problém eliminuje a detekuje významné body na povrchu polygonálnych modelov čeľustí. V metóde sú použité rôzne architektúry neurónových sietí, založené na architektúre U-Net. Viacpohľadový prístup presúva spracovanie do 2D, kde navrhnuté architektúry detekujú body regresiou tepelných máp z niekoľkých pohľadov. Pomocou konsezus metódy je následne z týchto pohľadov určená konečná pozícia bodov v 3D priestore. V práci sú predstavené experimenty s dvoma konsenzus metódami: stredná hodnota predikcií a geometrický prístup založený na algoritme RANSAC a metóde najmenších štvorcov. Experimenty ukázali, že varianta kombinujúca Attention U-Net, 100 pohľadov a geometrickú konsenus metódu je schopná detekovať významné body s chybou 1.20 +- 1.81 mm, pričom 94.01% predikcií dosahuje chybu menšiu ako 2 mm.
|
|
|
Watermarking 3D Models
Stehlík, Václav ; Beran, Vítězslav (referee) ; Španěl, Michal (advisor)
This bachelor's thesis is dedicated to embedding and extraction of watermarks to/from 3D models represented by polygonal mesh. The goal is to choose a suitable method for watermark embedding and extraction to/from a 3D model for the purpose of implementing a tool of such capabilities. The resulting solution uses methods proposed in article A Novel Blind Robust Digital Watermarking on 3D Meshes . Used methods create two watermarks. 'OTC' watermark that changes vertices of the model and 'Zero' watermark that does not change the model at all. These watermarks provide robustness against attacks and modification such as translation, scaling, rotation, simplification, random noise, remeshing and vertex reordering. Used methods also preserve imperceivability of watermark thanks to minimal changes of the model. Methods are blind and do not require original model. The tool is implemented in a form of simple web application which allows user to embed and extract watermark to/from files in stl format. This work includes testing and evaluation of success and usability of the tool against various attacks on chosen test set of models.
|
|
|
Analysis of Polygonal Models Using Neural Networks
Dronzeková, Michaela ; Zemčík, Pavel (referee) ; Kodym, Oldřich (advisor)
This thesis deals with rotation estimation of 3D model of human jaw. It describes and compares methods for direct analysis od 3D models as well as method to analyze model using rasterization. To evaluate perfomance of proposed method, a metric that computes number of cases when prediction was less than 30° from ground truth is used. Proposed method that uses rasterization, takes three x-ray views of model as an input and processes it with convolutional network. It achieves best preformance, 99% with described metric. Method to directly analyze polygonal model as a sequence uses attention mechanism to do so and was inspired by transformer architecture. A special pooling function was proposed for this network that decreases memory requirements of the network. This method achieves 88%, but does not use rasterization and can process polygonal model directly. It is not as good as rasterization method with x-ray display, byt it is better than rasterization method with model not rendered as x-ray. The last method uses graph representation of mesh. Graph network had problems with overfitting, that is why it did not get good results and I think this method is not very suitable for analyzing plygonal model.
|
|
|
Tooth Detection on Jaw 3D Computer Polygonal Model
Filip, Jan ; Šiler, Ondřej (referee) ; Kršek, Přemysl (advisor)
Aim of this bachelor thesis is a proposal and an implementation of a tool for the tooth detection on a 3D computer polygonal jaw model, which will allow to a user to obtain the shape of teeth on the jaw. This tool should to help in the stomatology. In the event of any inaccurary in the detection it will be possible to change certain parameters and methods, by which will be possible to influence the accuracy of detection. Program will import models in the STL format and subsequently will allow to the user to see the result of the detection in a browser. Teeth detection models will be possible to export single or together, in the STL format. The implementation language will be C++, compiler GNU C in distribution MinGW, toolkit to work with the 3D model will MDSTk and primarily its part VectorEntity. For imaging of thr 3D scene will be used OpenSceneGraph toolkit. Application will be developed primarily for MS Windows operating system but thanks to using GNU tools should be portable at the level of source codes to anothers operating systems.
|
|
|
Polygonal Models Decimation
Johannesová, Daniela ; Herout, Adam (referee) ; Kršek, Přemysl (advisor)
When representing objects with polygonal models, we must often use models consisting of large number of polygons. If we want to work in real time with such model the model must be simplificated. The process of simplification is called decimation. We know several decimation techniques for simplification of models. According to characteristics of model and target features of simplified model, we want to choose suitable method. So we need to know characteristics of particular methods and also their limits.
|
|
|
Deep Neural Networks for Landmark Detection on 3D Models
Kubík, Tibor ; Kodym, Oldřich (referee) ; Španěl, Michal (advisor)
Detekcia významných bodov je častým krokom pri analýze medicínskych dát. Čoraz bežnejšie sú tieto dáta reprezentované vo forme 3D modelov, príkladom sú povrchové skeny zubného oblúka pacienta. Hlboké neurónové siete sú vhodný spôsob, ako detekovať významné body v obraze. V prípade 3D dát je však toto spracovanie časovo i pamäťovo náročné, čo nevyhovuje požiadavkám kladeným medicínskymi aplikáciami. V tejto práci navrhujem metódu, ktorá tento problém eliminuje a detekuje významné body na povrchu polygonálnych modelov čeľustí. V metóde sú použité rôzne architektúry neurónových sietí, založené na architektúre U-Net. Viacpohľadový prístup presúva spracovanie do 2D, kde navrhnuté architektúry detekujú body regresiou tepelných máp z niekoľkých pohľadov. Pomocou konsezus metódy je následne z týchto pohľadov určená konečná pozícia bodov v 3D priestore. V práci sú predstavené experimenty s dvoma konsenzus metódami: stredná hodnota predikcií a geometrický prístup založený na algoritme RANSAC a metóde najmenších štvorcov. Experimenty ukázali, že varianta kombinujúca Attention U-Net, 100 pohľadov a geometrickú konsenus metódu je schopná detekovať významné body s chybou 1.20 +- 1.81 mm, pričom 94.01% predikcií dosahuje chybu menšiu ako 2 mm.
|
|
|
Analysis of Polygonal Models Using Neural Networks
Dronzeková, Michaela ; Zemčík, Pavel (referee) ; Kodym, Oldřich (advisor)
This thesis deals with rotation estimation of 3D model of human jaw. It describes and compares methods for direct analysis od 3D models as well as method to analyze model using rasterization. To evaluate perfomance of proposed method, a metric that computes number of cases when prediction was less than 30° from ground truth is used. Proposed method that uses rasterization, takes three x-ray views of model as an input and processes it with convolutional network. It achieves best preformance, 99% with described metric. Method to directly analyze polygonal model as a sequence uses attention mechanism to do so and was inspired by transformer architecture. A special pooling function was proposed for this network that decreases memory requirements of the network. This method achieves 88%, but does not use rasterization and can process polygonal model directly. It is not as good as rasterization method with x-ray display, byt it is better than rasterization method with model not rendered as x-ray. The last method uses graph representation of mesh. Graph network had problems with overfitting, that is why it did not get good results and I think this method is not very suitable for analyzing plygonal model.
|
|
|
Watermarking 3D Models
Stehlík, Václav ; Beran, Vítězslav (referee) ; Španěl, Michal (advisor)
This bachelor's thesis is dedicated to embedding and extraction of watermarks to/from 3D models represented by polygonal mesh. The goal is to choose a suitable method for watermark embedding and extraction to/from a 3D model for the purpose of implementing a tool of such capabilities. The resulting solution uses methods proposed in article A Novel Blind Robust Digital Watermarking on 3D Meshes . Used methods create two watermarks. 'OTC' watermark that changes vertices of the model and 'Zero' watermark that does not change the model at all. These watermarks provide robustness against attacks and modification such as translation, scaling, rotation, simplification, random noise, remeshing and vertex reordering. Used methods also preserve imperceivability of watermark thanks to minimal changes of the model. Methods are blind and do not require original model. The tool is implemented in a form of simple web application which allows user to embed and extract watermark to/from files in stl format. This work includes testing and evaluation of success and usability of the tool against various attacks on chosen test set of models.
|
guest ::
