Book spine segmentation for various book orientations

Chih Chang Yu, Rui Jun Zhang, Hsu Yung Cheng

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

Abstract

This work proposes a novel approach to separate books which are placed disorderly on a bookshelf. This work applies grayscale morphological reconstruction and Lo gradient minimization to erase texts on books while preserving book boundaries. Afterwards, Hough transform and the proposed SVM-based recovery scheme are applied to find book boundaries accurately. The proposed SVM-based recovery scheme is able to recover the book boundaries missed by Hough transform without incurring noticeable false alarm detections. Experiments demonstrate that the proposed method is able to segment books which are placed in various orientations.

Original languageEnglish
Title of host publication2015 IEEE 4th Global Conference on Consumer Electronics, GCCE 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages99-100
Number of pages2
ISBN (Electronic)9781479987511
DOIs
StatePublished - 3 Feb 2016
Event4th IEEE Global Conference on Consumer Electronics, GCCE 2015 - Osaka, Japan
Duration: 27 Oct 201530 Oct 2015

Publication series

Name2015 IEEE 4th Global Conference on Consumer Electronics, GCCE 2015

Conference

Conference4th IEEE Global Conference on Consumer Electronics, GCCE 2015
Country/TerritoryJapan
CityOsaka
Period27/10/1530/10/15

Keywords

  • L0 gradient minimization
  • Support Vector Machine
  • book spine segmentation
  • morphological reconstruction

Fingerprint

Dive into the research topics of 'Book spine segmentation for various book orientations'. Together they form a unique fingerprint.

Cite this