Presentation 2010-05-13
Pedestrian Location Estimation by First-Person View Camera
Yoshinari KAMEDA, Yuichi OHTA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a new pedestrian location method that exploits only a human-borne camera and runs on line. Pedestrian localization is an essential technology for pedestrian walking navigation. Unfortunately, commonly used GPS is not available, or unreliable in underground shopping mall, indoor paths, and urban streets with tall buildings. Our proposed method can estimate current position only by one image of the first person vision camera, assuming that a video that is taken by walking through the same path is given to our system in advance. As a first person camera is mounted on a body of a pedestrian, images taken by this camera always have weaving motion and obstacles such as other pedestrians. The proposed method adopts SURF based general image retrieval, but it exploits some conditions that are specific to first person vision camera so that it can produce reliable localization results. We take up one of the typical and difficult paths for experiment and evaluation; a path that starts from underground level of Tokyo station area to the ground level where tall buildings cover the sky. The result shows the proposed method can work well even in urban areas like there.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) First Person Vision / SLAM / Wearable camera / SURF / KD tree
Paper # IE2010-26,PRMU2010-14,MI2010-14
Date of Issue

Conference Information
Committee MI
Conference Date 2010/5/6(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Medical Imaging (MI)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Pedestrian Location Estimation by First-Person View Camera
Sub Title (in English)
Keyword(1) First Person Vision
Keyword(2) SLAM
Keyword(3) Wearable camera
Keyword(4) SURF
Keyword(5) KD tree
1st Author's Name Yoshinari KAMEDA
1st Author's Affiliation University of Tsukuba()
2nd Author's Name Yuichi OHTA
2nd Author's Affiliation University of Tsukuba
Date 2010-05-13
Paper # IE2010-26,PRMU2010-14,MI2010-14
Volume (vol) vol.110
Number (no) 28
Page pp.pp.-
#Pages 6
Date of Issue