Texture Synthesis with Recurrent Variational Auto-Encoder

2017/12/23 by Chandra, Rohan, Grover, Sachin, Lee, Kyungjun +2
#Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.1712.08838

Abstract

We propose a recurrent variational auto-encoder for texture synthesis. A novel loss function, FLTBNK, is used for training the texture synthesizer. It is rotational and partially color invariant loss function. Unlike L2 loss, FLTBNK explicitly models the correlation of color intensity between pixels. Our texture synthesizer generates neighboring tiles to expand a sample texture and is evaluated using various texture patterns from Describable Textures Dataset (DTD). We perform both quantitative and qualitative experiments with various loss functions to evaluate the performance of our proposed loss function (FLTBNK) --- a mini-human subject study is used for the qualitative evaluation.

Related

• Transposer: Universal Texture Synthesis Using Feature Maps as Transposed Convolution Filter
• Optimal Textures: Fast and Robust Texture Synthesis and Style Transfer through Optimal Transport
• Generating Non-Stationary Textures using Self-Rectification
• Texture synthesis via projection onto multiscale, multilayer statistics
• Synthesis of Near-regular Natural Textures