An automatic stranger removal photography system using panoramic inpainting

W. G.C.W. Kumara, Timothy K. Shih, Shih Ming Chang

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

Moving human objects (or 'strangers' as referred here) in the background obscure the scenery in regular panorama photography. Hence this paper proposes a novel method to remove the strangers from the background of the panorama. In the proposed system the object segmentation, dynamic background inpainting, and panorama creation phases are used to compose a panorama. If the detected face data of the stranger are not in the face database of the camera, the stranger is automatically removed from the panorama in the background repair phase which is known as inpainting. Panorama creation of the developed system is fully automatic. If a stranger in the background that the user intend to remove from the panorama is not automatically detected, the user can manually select that stranger in the object segmentation phase. Experimental results demonstrate that the proposed system effectively removes strangers in the background and generates the panorama. Complete results set can be accessed at http://video.minelab.tw/PanoramaInpainting.

Original languageEnglish
Pages (from-to)1919-1935
Number of pages17
JournalJournal of Information Science and Engineering
Volume31
Issue number6
StatePublished - Nov 2015

Keywords

  • Automatic grabcut
  • Face matching
  • Inpainting
  • Motion estimation
  • Panorama

Fingerprint

Dive into the research topics of 'An automatic stranger removal photography system using panoramic inpainting'. Together they form a unique fingerprint.

Cite this