On Demand Solid Texture Synthesis Using Deep 3D Networks

Jorge Gutierrez (1) , Julien Rabin (2) , Bruno Galerne (3) , Thomas Hurtut (1)

1 Polytechnique Montréal
2 Equipe Image - Laboratoire GREYC - UMR6072
3 IDP - Institut Denis Poisson

This paper describes a novel approach for on demand volumetric texture synthesis based on a deep learning framework that allows for generation of high quality 3D data at interactive rates. Based on a few example images of textures, a generative network is trained to synthesize coherent solid textures that reproduce the visual characteristics of the examples along some directions. To cope with memory limitations and computation complexity that are inherent to both high resolution and 3D processing on GPU, only 2D textures referred to as “slices” are generated during the training stage. These synthetic textures are compared to exemplar images via a perceptual loss function based on a pre-trained deep network. The proposed network is very light (less than 100k parameters), therefore it only requires sustainable training (i.e. few hours) and is capable of very fast generation (around a second for 256^3 voxels) on a single GPU. The synthesized volumes have good visual results that are at least equivalent to the state-of-the-art patch based approaches. They are naturally seamlessly tileable and can be fully generated in parallel.

Mots clés: Solid textures; Volumetric data generation; Texture synthesis; Convolutional neural networks; On demand texture synthesis; Generative networks; Deep learning

Domaines: Machine Learning [stat.ML]; Traitement des images [eess.IV]; Synthèse d'image et réalité virtuelle [cs.GR]

Jorge Gutierrez, Julien Rabin, Bruno Galerne, Thomas Hurtut. On Demand Solid Texture Synthesis Using Deep 3D Networks. 2018. ⟨hal-01678122v2⟩