LAEM

2000년 도입 · 논문 1편에서 사용

Method that adapts expectation-maximization (EM) to aggregate annotations that work on tasks with combined localization and classification like object detection and instance segmentation.

출처: Drawing the Same Bounding Box Twice? Coping Noisy Annotations in Object Detection with Repeated Labels

소개 논문: Drawing the Same Bounding Box Twice? Coping Noisy Annotations in Object Detection with Repeated Labels