DSLA: Dynamic smooth label assignment for efficient anchor-free object detection

作者：

Highlights：

• The inconsistencies of classification and quality estimation are analyzed. Dynamic smooth label assignment is proposed to address the problems.

• Interval relaxation strategy is proposed and combined with the improved centerness score. The assigned label is smoothed to a continuous value.

• IoU score is dynamically calculated and coupled with the smooth label, resulting in dynamic smooth label.

• DSLA is applied to popular anchor-free detectors. Comprehensive experiments are carried out on MS COCO to demonstrate the effectiveness.

摘要

•The inconsistencies of classification and quality estimation are analyzed. Dynamic smooth label assignment is proposed to address the problems.•Interval relaxation strategy is proposed and combined with the improved centerness score. The assigned label is smoothed to a continuous value.•IoU score is dynamically calculated and coupled with the smooth label, resulting in dynamic smooth label.•DSLA is applied to popular anchor-free detectors. Comprehensive experiments are carried out on MS COCO to demonstrate the effectiveness.

论文关键词：Convolutional neural network,Object detection,Centerness score,Intersection-of-union

论文评审过程：Received 14 February 2022, Revised 23 May 2022, Accepted 20 June 2022, Available online 22 June 2022, Version of Record 27 June 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108868