general
critique
bearish
Existing replay-based or regularization-based methods for domain-incremental change detection often fail to scale to long domain sequences, leading to knowledge degradation or increased computational cost
Computer Vision27 Jul 2026