BLNet: Boundary Points Localization Network for Object Detection

Jiaoyang An, Bo Ma

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Currently, almost all of the two-stage object detectors treat bounding box localization as an offset regression problem in the second stage. However, the spatial information in each Region of Interest (RoI) feature map is not considered in this pipeline. In this paper, we propose a novel boundary points localization network (BLNet) to predict the location of four boundary points (topmost, bottommost, leftmost, rightmost) of objects on RoI feature maps with a fully convolutional network. In addition, in order to compensate for the low resolution of the heatmaps, we use a differentiable operation called soft-argmax to convert the heatmaps into the numerical coordinates directly. Experiments on PASCAL VOC 2007 and 2012 datasets demonstrate that our BLNet significantly outperforms the traditional regression-based methods. Using ResNet-101 as the backbone, our method achieves 80.9% mAP on VOC 2007 and 78.7% mAP on VOC 2012 dataset.

Original languageEnglish
Title of host publication2020 IEEE 5th International Conference on Signal and Image Processing, ICSIP 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages281-285
Number of pages5
ISBN (Electronic)9781728168968
DOIs
Publication statusPublished - 23 Oct 2020
Event5th IEEE International Conference on Signal and Image Processing, ICSIP 2020 - Virtual, Nanjing, China
Duration: 23 Oct 202025 Oct 2020

Publication series

Name2020 IEEE 5th International Conference on Signal and Image Processing, ICSIP 2020

Conference

Conference5th IEEE International Conference on Signal and Image Processing, ICSIP 2020
Country/TerritoryChina
CityVirtual, Nanjing
Period23/10/2025/10/20

Keywords

  • CNN
  • deep learning
  • object detection
  • object localization

Fingerprint

Dive into the research topics of 'BLNet: Boundary Points Localization Network for Object Detection'. Together they form a unique fingerprint.

Cite this