search

JosephKJ / OWOD

(CVPR 2021 Oral) Open World Object Detection
https://josephkj.in
Apache License 2.0
1.02k stars 153 forks source link

Why Faster-RCNN trained on data and annotations of task 2 still can detect class of task 1? As shown in the red box. #89

Closed dongnana777 closed 2 years ago

dongnana777 commented 2 years ago

图片 1

JosephKJ commented 2 years ago

This is because of the learnings that it has had in the first step of learning (task 1). You can see that the performance is indeed very bad.