Revisiting the Generalization Problem of Low-level Vision Models Through the Lens of Image Deraining
This addresses the challenge of poor real-world performance for low-level vision models, offering incremental improvements to training strategies.
The paper tackles the generalization problem in low-level vision models by analyzing image deraining, revealing that overfitting to degradation patterns, not limited network capacity, is the key issue. It proposes strategies like balancing training data complexity and using content priors, showing improved generalization in deraining and denoising tasks.
Generalization remains a significant challenge for low-level vision models, which often struggle with unseen degradations in real-world scenarios despite their success in controlled benchmarks. In this paper, we revisit the generalization problem in low-level vision models. Image deraining is selected as a case study due to its well-defined and easily decoupled structure, allowing for more effective observation and analysis. Through comprehensive experiments, we reveal that the generalization issue is not primarily due to limited network capacity but rather the failure of existing training strategies, which leads networks to overfit specific degradation patterns. Our findings show that guiding networks to focus on learning the underlying image content, rather than the degradation patterns, is key to improving generalization. We demonstrate that balancing the complexity of background images and degradations in the training data helps networks better fit the image distribution. Furthermore, incorporating content priors from pre-trained generative models significantly enhances generalization. Experiments on both image deraining and image denoising validate the proposed strategies. We believe the insights and solutions will inspire further research and improve the generalization of low-level vision models.