Procedural Content Generation via Machine Learning in 2D Indoor Scene.

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F62690094%3A18450%2F20%3A50017135" target="_blank" >RIV/62690094:18450/20:50017135 - isvavai.cz</a>

Výsledek na webu

<a href="https://link.springer.com/chapter/10.1007/978-3-030-58465-8_3" target="_blank" >https://link.springer.com/chapter/10.1007/978-3-030-58465-8_3</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-030-58465-8_3" target="_blank" >10.1007/978-3-030-58465-8_3</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Procedural Content Generation via Machine Learning in 2D Indoor Scene.

Popis výsledku v původním jazyce

The article proposes a method of combining multiple deep forward neural networks to generate a distribution of objects in a 2D scene. The main concepts of machine learning, neural networks and procedural content generation concerning this intention are presented here. Additionally, these concepts are put into the context of computer graphics and used in a practical example of generating an indoor 2D scene. A method of vectorization of input datasets for training forward neural networks is proposed. Scene generation is based on the consequent placement of objects of different classes into the free space defining a room of a certain shape. Several evaluate methods have been proposed for testing the correctness of generation.

Název v anglickém jazyce

Procedural Content Generation via Machine Learning in 2D Indoor Scene.

Popis výsledku anglicky

The article proposes a method of combining multiple deep forward neural networks to generate a distribution of objects in a 2D scene. The main concepts of machine learning, neural networks and procedural content generation concerning this intention are presented here. Additionally, these concepts are put into the context of computer graphics and used in a practical example of generating an indoor 2D scene. A method of vectorization of input datasets for training forward neural networks is proposed. Scene generation is based on the consequent placement of objects of different classes into the free space defining a room of a certain shape. Several evaluate methods have been proposed for testing the correctness of generation.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2020

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Augmented Reality, Virtual Reality, and Computer Graphics. Part I

ISBN

978-3-030-58464-1

ISSN

0302-9743

e-ISSN

1611-3349

Počet stran výsledku

16

Strana od-do

34-49

Název nakladatele

Springer Nature Switzerland AG 2020

Místo vydání

Berlín

Místo konání akce

Lecce, Italy

Datum konání akce

7. 9. 2020

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000713422700003