The author presents MultIOD, a class-incremental object detector based on CenterNet, emphasizing the importance of rehearsal-free and anchor-free object detection.
MultIODは、Anchor-freeでRehearsal-freeなContinual Learningに基づくクラス増分オブジェクト検出器です。
MultIOD is a class-incremental object detector based on CenterNet, focusing on rehearsal-free and anchor-free object detection.
Inversed Objects Replay (IOR) efficiently generates old-class object samples to mitigate catastrophic forgetting in incremental object detection without co-occurrence of old and new class objects.
CASA leverages shared attributes in vision-language models to improve incremental object detection, addressing the challenge of background shift by efficiently transferring knowledge between tasks.