Corner Proposal Network for Anchor-free, Two-stage Object Detection

  • 2020-07-27 19:04:57
  • Kaiwen Duan, Lingxi Xie, Honggang Qi, Song Bai, Qingming Huang, Qi Tian
  • 24

Abstract

The goal of object detection is to determine the class and location ofobjects in an image. This paper proposes a novel anchor-free, two-stageframework which first extracts a number of object proposals by findingpotential corner keypoint combinations and then assigns a class label to eachproposal by a standalone classification stage. We demonstrate that these twostages are effective solutions for improving recall and precision,respectively, and they can be integrated into an end-to-end network. Ourapproach, dubbed Corner Proposal Network (CPN), enjoys the ability to detectobjects of various scales and also avoids being confused by a large number offalse-positive proposals. On the MS-COCO dataset, CPN achieves an AP of 49.2%which is competitive among state-of-the-art object detection methods. CPN alsofits the scenario of computational efficiency, which achieves an AP of41.6%/39.7% at 26.2/43.3 FPS, surpassing most competitors with the sameinference speed. Code is available at https://github.com/Duankaiwen/CPNDet

 

Quick Read (beta)

loading the full paper ...