A Generative Adversarial Framework for Optimizing Image Matting and Harmonization Simultaneously

TL;DR

This paper introduces a GAN-based framework that jointly optimizes image matting and harmonization tasks, leading to more natural composite images by leveraging a self-attention discriminator.

Contribution

The novel approach simultaneously optimizes matting and harmonization using a self-attention GAN, improving performance over separate optimization methods.

Findings

Enhanced image quality in composite results

Effective joint optimization of matting and harmonization

Validated on a new dataset with promising results

Abstract

Image matting and image harmonization are two important tasks in image composition. Image matting, aiming to achieve foreground boundary details, and image harmonization, aiming to make the background compatible with the foreground, are both promising yet challenging tasks. Previous works consider optimizing these two tasks separately, which may lead to a sub-optimal solution. We propose to optimize matting and harmonization simultaneously to get better performance on both the two tasks and achieve more natural results. We propose a new Generative Adversarial (GAN) framework which optimizing the matting network and the harmonization network based on a self-attention discriminator. The discriminator is required to distinguish the natural images from different types of fake synthesis images. Extensive experiments on our constructed dataset demonstrate the effectiveness of our proposed method. Our dataset and dataset generating pipeline can be found in \url{https://git.io/HaMaGAN}