Mirror-Yolo: A Novel Attention Focus, Instance Segmentation and Mirror Detection Model

Fengze Li, Jieming Ma*, Zhongbei Tian, Ji Ge, Hai Ning Liang, Yungang Zhang, Tianxi Wen

*Corresponding author for this work

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

2 Citations (Scopus)

Abstract

Mirrors can degrade the performance of computer vision models, but research into detecting them is in the preliminary phase. YOLOv4 achieves phenomenal results in terms of object detection accuracy and speed, but it still fails in detecting mirrors. Thus, we propose Mirror-YOLO, which targets mirror detection, containing a novel attention focus mechanism for features acquisition, a hypercolumn-stairstep approach to better fusion the feature maps, and the mirror bounding polygons for instance segmentation. Compared to the existing mirror detection networks and YOLO series, our proposed network achieves superior performance in average accuracy on our proposed mirror dataset and another state-of-art mirror dataset, which demonstrates the validity and effectiveness of Mirror-YOLO.

Original languageEnglish
Title of host publication2022 7th International Conference on Frontiers of Signal Processing, ICFSP 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages76-80
Number of pages5
ISBN (Electronic)9781665481588
DOIs
Publication statusPublished - 2022
Event7th International Conference on Frontiers of Signal Processing, ICFSP 2022 - Paris, France
Duration: 7 Sept 20229 Sept 2022

Publication series

Name2022 7th International Conference on Frontiers of Signal Processing, ICFSP 2022

Conference

Conference7th International Conference on Frontiers of Signal Processing, ICFSP 2022
Country/TerritoryFrance
CityParis
Period7/09/229/09/22

Keywords

  • Object detection
  • YOLOv4
  • attention mechanism
  • mirror bounding polygons
  • mirror detection

Fingerprint

Dive into the research topics of 'Mirror-Yolo: A Novel Attention Focus, Instance Segmentation and Mirror Detection Model'. Together they form a unique fingerprint.

Cite this