Previous [ 1] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] [ 8] [ 9] [ 10] [ 11] [ 12] [ 13] [ 14] [ 15] [ 16] [ 17]

Journal of Information Science and Engineering, Vol. 31 No. 6, pp. 1919-1935 (November 2015)


An Automatic Stranger Removal Photography System Using Panoramic Inpainting


W. G. C. W. KUMARA*, TIMOTHY K. SHIH* AND SHIH-MING CHANG+
* Department of Computer Science and Information Engineering
National Central University
Jhongli City, Taoyuan, 320 Taiwan
+ Department of Computer Science and Information Engineering
Tamkang University
Tamsui, New Taipei City, 251 Taiwan
E-mail: {chinthakawk; timothykshih}@gmail.com; rest306@hotmail.com

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.

Keywords: panorama, inpainting, automatic grabcut, motion estimation, face matching

Full Text (•Ģ§Śņ…) Retrieve PDF document (201511_06.pdf)

Received June 25, 2014; revised September 4, 2014 & January 1&27, 2015; accepted February 4, 2015.
Communicated by Chu-Song Chen.
1 http://www.cs.bath.ac.uk/brown/autostitch/autostitch.html