Texture Synthesis with Recurrent Variational Auto-Encoder

TL;DR

This paper introduces a recurrent variational auto-encoder with a novel FLTBNK loss function for texture synthesis, capable of generating seamless textures with improved color correlation modeling, evaluated through quantitative, qualitative, and human studies.

Contribution

It presents a new recurrent variational auto-encoder architecture combined with the FLTBNK loss function for enhanced texture synthesis.

Findings

FLTBNK outperforms L2 loss in modeling color correlations.

The synthesizer generates seamless neighboring tiles.

Qualitative and human studies validate the effectiveness of FLTBNK.

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.