AccLoc: Anchor-Free and two-stage detector for accurate object localization
作者：
Highlights：
• We proposed an anchor-free, NMS-free and two-stage object detection framework for accurate localization.
• Adaptive proposal jitter strategy contributes to learning the mappings between the input regional features and the desired output offsets.
• Proposal fine-tuning strategy encourages the networks to cover various and helpful features for the subsequent bounding box regression process.
• Extensive experiments on COCO, KITTI, and BDD100K were conducted to demonstrate the effectiveness of our proposed approach.
摘要
•We proposed an anchor-free, NMS-free and two-stage object detection framework for accurate localization.•Adaptive proposal jitter strategy contributes to learning the mappings between the input regional features and the desired output offsets.•Proposal fine-tuning strategy encourages the networks to cover various and helpful features for the subsequent bounding box regression process.•Extensive experiments on COCO, KITTI, and BDD100K were conducted to demonstrate the effectiveness of our proposed approach.
论文关键词：Object detection,Accurate localization,Anchor-free,NMS-free,Two-stage
论文评审过程：Received 29 May 2021, Revised 1 December 2021, Accepted 5 January 2022, Available online 8 January 2022, Version of Record 12 February 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108523
