.png){kind=icon}
.png){kind=icon}
Return
DBIDM: Implementing Blind Image Separation through a Dual Branch Interactive Diffusion Model
.png){kind=icon}
PRE
AI
DOI：10.1016/j.patrec.2025.11.038.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Wepropose a novel framework named DBIDM. To the best of our knowledge, this is the first study to apply diffusion models to the task of blind image separation. • Within the reverse denoising process of our framework, we integrate a specifically designed Wavelet Interactive Decou pling Module to cooperatively separate source images from dual domains. • Extensive experiments conducted on constructed datasets for deraining, desnowing, and complex mixtures demonstrate that our method achieves superior performance in blind image separation compared to other approaches based on Trans former and GAN models.