Deep Image Harmonization with Globally Guided Feature Transformation and Relation Distillation

Li Niu, Linfeng Tan, Xinhao Tao, Junyan Cao, Fengjun Guo, Teng Long, Liqing Zhang; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023, pp. 7723-7732

Abstract

Given a composite image, image harmonization aims to adjust the foreground illumination to be consistent with background. Previous methods have explored transforming foreground features to achieve competitive performance. In this work, we show that using global information to guide foreground feature transformation could achieve significant improvement. Besides, we propose to transfer the foreground-background relation from real images to composite images, which can provide intermediate supervision for the transformed encoder features. Additionally, considering the drawbacks of existing harmonization datasets, we also contribute a ccHarmony dataset which simulates the natural illumination variation. Extensive experiments on iHarmony4 and our contributed dataset demonstrate the superiority of our method. Our ccHarmony dataset is released at https://github.com/bcmi/Image-Harmonization-Dataset-ccHarmony.

Related Material

[pdf] [supp] [arXiv]
[bibtex]
@InProceedings{Niu_2023_ICCV, author = {Niu, Li and Tan, Linfeng and Tao, Xinhao and Cao, Junyan and Guo, Fengjun and Long, Teng and Zhang, Liqing}, title = {Deep Image Harmonization with Globally Guided Feature Transformation and Relation Distillation}, booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)}, month = {October}, year = {2023}, pages = {7723-7732} }