Large Scale Image Completion via Co-Modulated Generative Adversarial Networks

Numerous task-specific variants of conditional generative adversarial networks have been developed for image completion. Yet, a serious limitation remains that all existing algorithms tend to fail when handling large-scale missing regions. To overcome this challenge, we propose a generic new approach that bridges the gap between image-conditional and recent modulated unconditional generative architectures via co-modulation of both conditional and stochastic style representations. Also, due to the lack of good quantitative metrics for image completion, we propose the new Paired/Unpaired Inception Discriminative Score (P-IDS/U-IDS), which robustly measures the perceptual fidelity of inpainted images compared to real images via linear separability in a feature space. Experiments demonstrate superior performance in terms of both quality and diversity over state-of-the-art methods in free-form image completion and easy generalization to image-to-image translation. Code is available at

PDF Abstract ICLR 2021 PDF ICLR 2021 Abstract

Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Inpainting CelebA-HQ CoModGAN FID 5.65 # 3
P-IDS 11.23 # 2
U-IDS 22.54 # 2
Image Inpainting FFHQ 512 x 512 CoModGAN P-IDS 16.6% # 1
U-IDS 29.4% # 1
FID 3.7 # 3
Image Inpainting Places2 CoModGAN FID 2.92 # 3
P-IDS 19.64 # 3
U-IDS 35.78 # 3


No methods listed for this paper. Add relevant methods here