Timezone: »

 
Poster
Towards Coherent Image Inpainting Using Denoising Diffusion Implicit Models
Guanhua Zhang · Jiabao Ji · Yang Zhang · Mo Yu · Tommi Jaakkola · Shiyu Chang

Thu Jul 27 01:30 PM -- 03:00 PM (PDT) @ Exhibit Hall 1 #632
in Poster Session 5 »
Event URL: https://github.com/UCSB-NLP-Chang/CoPaint »

Image inpainting refers to the task of generating a complete, natural image based on a partially revealed reference image. Recently, many research interests have been focused on addressing this problem using fixed diffusion models. These approaches typically directly replace the revealed region of the intermediate or final generated images with that of the reference image or its variants. However, since the unrevealed regions are not directly modified to match the context, it results in incoherence between revealed and unrevealed regions. To address the incoherence problem, a small number of methods introduce a rigorous Bayesian framework, but they tend to introduce mismatches between the generated and the reference images due to the approximation errors in computing the posterior distributions. In this paper, we propose CoPaint, which can coherently inpaint the whole image without introducing mismatches. CoPaint also uses the Bayesian framework to jointly modify both revealed and unrevealed regions but approximates the posterior distribution in a way that allows the errors to gradually drop to zero throughout the denoising steps, thus strongly penalizing any mismatches with the reference image. Our experiments verify that CoPaint can outperform the existing diffusion-based methods under both objective and subjective metrics.

[ PDF]  [ Poster] 

Author Information

Guanhua Zhang (UC, Santa Barbara)
Jiabao Ji (University of California, Santa Barbara)
Yang Zhang (MIT-IBM Watson AI Lab)
Mo Yu (IBM Research)
Tommi Jaakkola (MIT)
Shiyu Chang (UCSB)

More from the Same Authors