MineGAN plus plus : Mining Generative Models for Efficient Knowledge Transfer to Limited Data DomainsShow others and affiliations
2023 (English)In: International Journal of Computer Vision, ISSN 0920-5691, E-ISSN 1573-1405, Vol. 132, no 2, p. 490-514Article in journal (Refereed) Published
Abstract [en]
Given the often enormous effort required to train GANs, both computationally as well as in dataset collection, the re-use of pretrained GANs largely increases the potential impact of generative models. Therefore, we propose a novel knowledge transfer method for generative models based on mining the knowledge that is most beneficial to a specific target domain, either from a single or multiple pretrained GANs. This is done using a miner network that identifies which part of the generative distribution of each pretrained GAN outputs samples closest to the target domain. Mining effectively steers GAN sampling towards suitable regions of the latent space, which facilitates the posterior finetuning and avoids pathologies of other methods, such as mode collapse and lack of flexibility. Furthermore, to prevent overfitting on small target domains, we introduce sparse subnetwork selection, that restricts the set of trainable neurons to those that are relevant for the target dataset. We perform comprehensive experiments on several challenging datasets using various GAN architectures (BigGAN, Progressive GAN, and StyleGAN) and show that the proposed method, called MineGAN, effectively transfers knowledge to domains with few target images, outperforming existing methods. In addition, MineGAN can successfully transfer knowledge from multiple pretrained GANs. MineGAN.
Place, publisher, year, edition, pages
SPRINGER , 2023. Vol. 132, no 2, p. 490-514
National Category
Computer Sciences
Identifiers
URN: urn:nbn:se:liu:diva-198074DOI: 10.1007/s11263-023-01882-yISI: 001062331300001OAI: oai:DiVA.org:liu-198074DiVA, id: diva2:1800106
Note
Funding Agencies|Huawei Kirin Solution; MCIN/AEI [PID2019-104174GB-I00, PID2021-128178OB-I00]; ERDF A way of making Europe; Ramon y Cajal fellowship - MCIN/AEI [RYC2019-027020-I]; CERCA Programme of Generalitat de Catalunya; Youth Foundation [62202243]
2023-09-252023-09-252024-04-12