# Terrain Generation With Deep Learning | Two Minute Papers #208 ## Метаданные - **Канал:** Two Minute Papers - **YouTube:** https://www.youtube.com/watch?v=NEscK5RCtlo - **Дата:** 22.11.2017 - **Длительность:** 3:34 - **Просмотры:** 78,733 ## Описание The paper "Interactive Example-Based Terrain Authoring with Conditional Generative Adversarial Networks" is available here: https://hal.archives-ouvertes.fr/hal-01583706/file/tog.pdf We would like to thank our generous Patreon supporters who make Two Minute Papers possible: Andrew Melnychuk, Brian Gilman, Christian Ahlin, Christoph Jadanowski, Dave Rushton-Smith, Dennis Abts, Eric Haddad, Esa Turkulainen, Evan Breznyik, Kaben Gabriel Nanlohy, Malek Cellier, Marten Rauschenberg, Michael Albrecht, Michael Jensen, Michael Orenstein, Raul Araújo da Silva, Robin Graham, Steef, Steve Messina, Sunil Kim, Torsten Reil. https://www.patreon.com/TwoMinutePapers Two Minute Papers Merch: US: http://twominutepapers.com/ EU/Worldwide: https://shop.spreadshirt.net/TwoMinutePapers/ Music: Antarctica by Audionautix is licensed under a Creative Commons Attribution license (https://creativecommons.org/licenses/by/4.0/) Artist: http://audionautix.com/ Splash screen/thumbnail design: Felícia Fehér - http://felicia.hu Károly Zsolnai-Fehér's links: Facebook: https://www.facebook.com/TwoMinutePapers/ Twitter: https://twitter.com/karoly_zsolnai Web: https://cg.tuwien.ac.at/~zsolnai/ ## Содержание ### [0:00](https://www.youtube.com/watch?v=NEscK5RCtlo) Dear Fellow Scholars, this is Two Minute Papers with Károly Zsolnai-Fehér. We have recently witnessed the emergence of neural network-based techniques that are able to synthesize all sorts of images. Our previous episode was about NVIDIA's algorithm that created high resolution images of imaginary celebrities that was a really cool application of Generative Adversarial Networks. This architecture means that we have a pair of neural networks, one that learns to generate new images, and the other learns to tell a fake image from a real one. As they compete against each other, they get better and better without any human interaction. So we can clearly use them to create 2D images, but why stop there? Why not use this technique, for instance, to create assets for digital media? So instead of 2D images, let's try to adapt these networks to generate high-resolution 3D models of terrains that we can use to populate a virtual world. Both computer games and the motion picture industry could benefit greatly from such a tool. This process is typically done via procedural generation, which is basically a sort of guided random terrain generation. Here, we can have a more direct effect on the output without putting in tens of hours of work to get the job done. In the first training step, this technique learns how an image of a terrain corresponds to input drawings. Then, we will be able to sketch a draft of a landscape with rivers, ridges, valleys and the algorithm will output a high quality model of the terrain itself. During this process, we can have a look at the current output and refine our drawings in the meantime, leading to a super efficient process where we can go from a thought to a high-quality final result within a few seconds without being bogged down with the technical details. What's more, it can also not only deal with erased subregions, but it can also automatically fill them with sensible information to save time for us. What an outstanding convenience feature! And, the algorithm can also perform physical manipulations like erosion to the final results. After the training for the erosion step is done, the computational cost is practically zero, for instance, running an erosion simulator on this piece of data would take around 40 ### [2:13](https://www.youtube.com/watch?v=NEscK5RCtlo&t=133s) Authoring session (Untrained user) seconds, where the neural network can do it in 25 milliseconds. The full simulation would almost be a minute, where the network can mimic its results practically instantaneously. A limitation of this technique is that if the input is too sparse, unpleasant grid artifacts may appear. There are tons of more cool features in the paper, make sure to have a look, as always ### [2:39](https://www.youtube.com/watch?v=NEscK5RCtlo&t=159s) Authoring session (Expert artist) it is available in the video description. This is a really well thought out and well-presented work that I expect to be a true powerhouse ### [2:46](https://www.youtube.com/watch?v=NEscK5RCtlo&t=166s) Lambda sketch for terrain authoring in the future. And, in the meantime, we have reached a hundred thousand subscribers. A hundred thousand Fellow Scholars. Wow. This is absolutely amazing and honestly I never thought that this would ever happen. ### [2:59](https://www.youtube.com/watch?v=NEscK5RCtlo&t=179s) Siggraph levelset sketch So, happy paperversary, thank you very much for coming along on this journey of science and I am very happy to see that the series brings joy and learning to more people than ### [3:08](https://www.youtube.com/watch?v=NEscK5RCtlo&t=188s) Island ever. Thanks for watching and for your generous support, and I'll see you next time! --- *Источник: https://ekstraktznaniy.ru/video/14552*