TY - GEN
T1 - GSDD
T2 - 38th AAAI Conference on Artificial Intelligence, AAAI 2024
AU - Zhang, Haiyu
AU - Su, Shaolin
AU - Zhu, Yu
AU - Sun, Jinqiu
AU - Zhang, Yanning
N1 - Publisher Copyright:
Copyright © 2024, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.
PY - 2024/3/25
Y1 - 2024/3/25
N2 - Single image super-resolution (SISR), especially in the real world, usually builds a large amount of LR-HR image pairs to learn representations that contain rich textural and structural information. However, relying on massive data for model training not only reduces training efficiency, but also causes heavy data storage burdens. In this paper, we attempt a pioneering study on dataset distillation (DD) for SISR problems to explore how data could be slimmed and compressed for the task. Unlike previous coreset selection methods which select a few typical examples directly from the original data, we remove the limitation that the selected data cannot be further edited, and propose to synthesize and optimize samples to preserve more task-useful representations. Concretely, by utilizing pre-trained GANs as a suitable approximation of realistic data distribution, we propose GSDD, which distills data in a latent generative space based on GAN-inversion techniques. By optimizing them to match with the practical data distribution in an informative feature space, the distilled data could then be synthesized. Experimental results demonstrate that when trained with our distilled data, GSDD can achieve comparable performance to the state-of-the-art (SOTA) SISR algorithms, while a nearly ×8 increase in training efficiency and a saving of almost 93.2% data storage space can be realized. Further experiments on challenging real-world data also demonstrate the promising generalization ability of GSDD.
AB - Single image super-resolution (SISR), especially in the real world, usually builds a large amount of LR-HR image pairs to learn representations that contain rich textural and structural information. However, relying on massive data for model training not only reduces training efficiency, but also causes heavy data storage burdens. In this paper, we attempt a pioneering study on dataset distillation (DD) for SISR problems to explore how data could be slimmed and compressed for the task. Unlike previous coreset selection methods which select a few typical examples directly from the original data, we remove the limitation that the selected data cannot be further edited, and propose to synthesize and optimize samples to preserve more task-useful representations. Concretely, by utilizing pre-trained GANs as a suitable approximation of realistic data distribution, we propose GSDD, which distills data in a latent generative space based on GAN-inversion techniques. By optimizing them to match with the practical data distribution in an informative feature space, the distilled data could then be synthesized. Experimental results demonstrate that when trained with our distilled data, GSDD can achieve comparable performance to the state-of-the-art (SOTA) SISR algorithms, while a nearly ×8 increase in training efficiency and a saving of almost 93.2% data storage space can be realized. Further experiments on challenging real-world data also demonstrate the promising generalization ability of GSDD.
UR - http://www.scopus.com/inward/record.url?scp=85189529111&partnerID=8YFLogxK
U2 - 10.1609/aaai.v38i7.28534
DO - 10.1609/aaai.v38i7.28534
M3 - 会议稿件
AN - SCOPUS:85189529111
T3 - Proceedings of the AAAI Conference on Artificial Intelligence
SP - 7069
EP - 7077
BT - Technical Tracks 14
A2 - Wooldridge, Michael
A2 - Dy, Jennifer
A2 - Natarajan, Sriraam
PB - Association for the Advancement of Artificial Intelligence
Y2 - 20 February 2024 through 27 February 2024
ER -