Learning with Style: Continual Semantic Segmentation Across Tasks and Domains
This work addresses a critical challenge for real-world AI systems like autonomous vehicles, which must adapt to varying tasks and domains, though it is incremental in combining existing techniques.
The paper tackles the unified problem of continual semantic segmentation across both tasks and domains, addressing semantic shifts in input and label spaces, and shows that the proposed Learning with Style (LwS) framework outperforms existing approaches on autonomous driving datasets.
Deep learning models dealing with image understanding in real-world settings must be able to adapt to a wide variety of tasks across different domains. Domain adaptation and class incremental learning deal with domain and task variability separately, whereas their unified solution is still an open problem. We tackle both facets of the problem together, taking into account the semantic shift within both input and label spaces. We start by formally introducing continual learning under task and domain shift. Then, we address the proposed setup by using style transfer techniques to extend knowledge across domains when learning incremental tasks and a robust distillation framework to effectively recollect task knowledge under incremental domain shift. The devised framework (LwS, Learning with Style) is able to generalize incrementally acquired task knowledge across all the domains encountered, proving to be robust against catastrophic forgetting. Extensive experimental evaluation on multiple autonomous driving datasets shows how the proposed method outperforms existing approaches, which prove to be ill-equipped to deal with continual semantic segmentation under both task and domain shift.