Object Detection Based on Sparse Representation of Foreground

Zhenyue Zhu, Shujing Lyu, Xiao Tu, Yue Lu

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

1 Scopus citations

Abstract

Objects detection can be regard as the segmentation of foreground from background. In this paper, we propose a foreground segmentation method based on sparse representation of direction features for threat object detection in X-ray images. The threat objects are supposed as foreground and all other contents in the images are background. We extract the direction features to make up foreground dictionary firstly. Then we search the foreground area in the test image through sparse representation of their direction features by foreground dictionary. The experimental results show that this proposed method is robust to the X-ray images with different backgrounds.

Original languageEnglish
Title of host publicationPattern Recognition and Artificial Intelligence - International Conference, ICPRAI 2020, Proceedings
EditorsYue Lu, Nicole Vincent, Pong Chi Yuen, Wei-Shi Zheng, Farida Cheriet, Ching Y. Suen
PublisherSpringer Science and Business Media Deutschland GmbH
Pages465-473
Number of pages9
ISBN (Print)9783030598297
DOIs
StatePublished - 2020
Event2nd International Conference on Pattern Recognition and Artificial Intelligence, ICPRAI 2020 - Zhongshan, China
Duration: 19 Oct 202023 Oct 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12068 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Conference on Pattern Recognition and Artificial Intelligence, ICPRAI 2020
Country/TerritoryChina
CityZhongshan
Period19/10/2023/10/20

Keywords

  • Object detection
  • Sparse representation
  • X-ray image

Fingerprint

Dive into the research topics of 'Object Detection Based on Sparse Representation of Foreground'. Together they form a unique fingerprint.

Cite this