Bounding Box Stability against Feature Dropout Reflects Detector Generalization across Environments

arXiv:2403.13803v1 Announce Type: new
Abstract: Bounding boxes uniquely characterize object detection, where a good detector gives accurate bounding boxes of categories of interest. However, in the real-world where test ground truths are not provided, it is non-trivial to find out whether bounding boxes are accurate, thus preventing us from assessing the detector generalization ability. In this work, we find under feature map dropout, good detectors tend to output bounding boxes whose locations do not change much, while bounding boxes of …

arxiv box cs.cv dropout environments feature stability type