Domain Generalization in Vision: A Survey

By Kaiyang Zhou, Ziwei Liu, Yu Qiao, Tao A. Xiang, Chen Change Loy. 2021

Subfields: Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Chen Change Loy
2. Luc Van Gool
3. Ziwei Liu
4. Xiaogang Wang
5. Ming-Hsuan Yang
6. Dacheng Tao
7. Yu Qiao
8. Hongsheng Li

In Representation & Self-Supervised Learning

1. Ziwei Liu
2. Jingkang Yang
3. Kaiyang Zhou
4. Chen Change Loy
5. Peidong Jia
6. Ji Qi
7. Bo Li
8. Tao Qin

Citations in the corpus

• Learning to Prompt for Vision-Language Models (2021, positive): “Since CoOp requires training on a specific data distribution, it risks learning spurious correlations that are detrimental to generalization in unseen distributions (domains), as suggested in recent studies (Taori et al., 2020; Zhou et al., 2021).”
• Conditional Prompt Learning for Vision-Language Models (2022, background): “Generalization to out-of-distribution data is a capability essential for machine learning models to succeed in practical applications [47, 61].”
• Self-Supervised Remote Sensing Feature Learning: Learning Paradigms, Challenges, and Future Works (2022, background): “supervised learning trains the model on a limited number of samples, resulting in the contradiction between the closed sample set and the dynamics and complexity of ground object features, which causes the performance collapse of the model [48]–[50].”
• ViDA: Homeostatic Visual Domain Adapter for Continual Test Time Adaptation (2023, methodology): “To investigate the domain generalization (DG) ability of our method, we follow the leave-one-domainout rule [65, 39] to leverage 10/15 domains of ImageNet-C as source domains for model training while the rest (5/15 domains) are treated as target domains without any form of adaptation.”
• Energy-Based Open-World Uncertainty Modeling for Confidence Calibration (2021, background): “A recent work [38] suggests that calibrating confidence across multiple domains is beneficial to OOD generalization [42].”
• Generalizing to Unseen Domains: A Survey on Domain Generalization (2021, background): “[3] also wrote a survey on DG, while their focus is in computer vision field.”