Written by Ying Qu ada.ying.qu@outlook.com This code is a demo code for our paper “Non-local Representation based Mutual Affine-Transfer Network for Photorealistic Stylization”, TPAMI 2021 The code is for research purpose only. All Rights Reserved

Usage

run "0_style_transfer" to perform style transfer with a single layer stick-breaking structure or "0_style_transfer_stack" for stacked layers.

Please adjust the parameters based on the input images.

The results of the proposed methods are located at the folders [data1_our_results_NL-MAT] and [data2_our_results_NL-MAT].

If you find the code helpful, please kindly cite the following paper.

Ying Qu, Zhenzhou Shao and Hairong Qi. “Non-local Representation based Mutual Affine-Transfer Network for Photorealistic Stylization”, IEEE Transactions on Pattern Analysis and Machine Intelligence，vol. 44, no. 10, pp. 7046-7061, 1 Oct. 2022, doi: 10.1109/TPAMI.2021.3095948.