Latent Feature-Guided Diffusion Models for Shadow Removal
This addresses shadow removal in images, which is important for computer vision applications, and is incremental as it builds on diffusion models with novel conditioning and training improvements.
The paper tackles the problem of recovering textures under shadows by proposing a diffusion model conditioned on a learned latent feature space, which outperforms the previous best method by 13% in RMSE on the AISTD dataset and by 82% on the DESOBA dataset for instance-level removal.
Recovering textures under shadows has remained a challenging problem due to the difficulty of inferring shadow-free scenes from shadow images. In this paper, we propose the use of diffusion models as they offer a promising approach to gradually refine the details of shadow regions during the diffusion process. Our method improves this process by conditioning on a learned latent feature space that inherits the characteristics of shadow-free images, thus avoiding the limitation of conventional methods that condition on degraded images only. Additionally, we propose to alleviate potential local optima during training by fusing noise features with the diffusion network. We demonstrate the effectiveness of our approach which outperforms the previous best method by 13% in terms of RMSE on the AISTD dataset. Further, we explore instance-level shadow removal, where our model outperforms the previous best method by 82% in terms of RMSE on the DESOBA dataset.