arXiv 2207.02039
PKD: General Distillation Framework for Object Detectors via Pearson Correlation Coefficient
By Weihan Cao, Yifan Zhang, et al.
Published 2022-07-05
Citation lineage
Review the prior work and downstream research connected to this paper.
Knowledge distillation(KD) is a widely-used technique to train compact models in object detection. However, there is still a lack of study on how to distill between heterogeneous detectors. In this paper, we empirically find that better FPN features from a heterogeneous teacher detector can help the student although their detection heads and label assignments are different. However, directly aligning the feature map…