Few Shot Generative Model Adaption via Relaxed Spatial Structural Alignment

Training a generative adversarial network (GAN) with limited data has been a challenging task. A feasible solution is to start with a GAN well-trained on a large scale source domain and adapt it to the target domain with a few samples, termed as few shot generative model adaption. However, existing...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online) pp. 11194 - 11203
Main Authors: Xiao, Jiayu, Li, Liang, Wang, Chaofei, Zha, Zheng-Jun, Huang, Qingming
Format: Conference Proceeding
Language:English
Published: IEEE 01.06.2022
Subjects:
Adaptation models
Codes
Computational modeling
Computer vision
Correlation
Image and video synthesis and generation; Transfer/low-shot/long-tail learning
Image coding
Training
ISSN:1063-6919
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Training a generative adversarial network (GAN) with limited data has been a challenging task. A feasible solution is to start with a GAN well-trained on a large scale source domain and adapt it to the target domain with a few samples, termed as few shot generative model adaption. However, existing methods are prone to model overfitting and collapse in extremely few shot setting (less than 10). To solve this problem, we propose a relaxed spatial structural alignment (RSSA) method to calibrate the target generative models during the adaption. We design a cross-domain spatial structural consistency loss comprising the self-correlation and disturbance correlation consistency loss. It helps align the spatial structural information between the synthesis image pairs of the source and target domains. To relax the cross-domain alignment, we compress the original latent space of generative models to a subspace. Image pairs generated from the subspace are pulled closer. Qualitative and quantitative experiments show that our method consistently surpasses the state-of-the-art methods in few shot setting. Our source code: https://github.com/StevenShaw1999/RSSA.
ISSN:1063-6919
DOI:10.1109/CVPR52688.2022.01092