Block Shuffle: A Method for High-resolution Fast Style Transfer with Limited Memory

TL;DR

The paper introduces 'block shuffle', a novel method that enables high-resolution fast style transfer on devices with limited memory by dividing the task into smaller, manageable subtasks, improving quality over existing feathering methods.

Contribution

It proposes a plug-in block shuffle technique that reduces memory usage for high-res style transfer without altering network architecture.

Findings

Outperforms feathering-based methods in image quality.

Enables high-resolution stylization on limited-memory devices.

Increases processing time but significantly reduces memory requirements.

Abstract

Fast Style Transfer is a series of Neural Style Transfer algorithms that use feed-forward neural networks to render input images. Because of the high dimension of the output layer, these networks require much memory for computation. Therefore, for high-resolution images, most mobile devices and personal computers cannot stylize them, which greatly limits the application scenarios of Fast Style Transfer. At present, the two existing solutions are purchasing more memory and using the feathering-based method, but the former requires additional cost, and the latter has poor image quality. To solve this problem, we propose a novel image synthesis method named \emph{block shuffle}, which converts a single task with high memory consumption to multiple subtasks with low memory consumption. This method can act as a plug-in for Fast Style Transfer without any modification to the network architecture. We use the most popular Fast Style Transfer repository on GitHub as the baseline. Experiments show that the quality of high-resolution images generated by our method is better than that of the feathering-based method. Although our method is an order of magnitude slower than the baseline, it can stylize high-resolution images with limited memory, which is impossible with the baseline. The code and models will be made available on \url{https://github.com/czczup/block-shuffle}.