Presentation 2019-11-06
[Poster Presentation] Performance Evaluations of Human Detection Method using 360-degree Image toward Sidewalk Congestion Estimation
Kazuki Murayama, Kenji Kanai, Masaru Takeuchi, Jiro Katto,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we evaluate the performance of human detection method using 360-degree images to estimate the sidewalk congestion level. Because the 360-degree image is captured by an omnidirectional camera, performed the image stitching, and transformed by equirectangular projection, the captured 360-degree image is distorted. Therefore, to detect humans with higher accuracy and longer distance, we do not use the distorted 360-degree image but use the “corrected” image by projection method. In addition, to maintain an advantage of high resolution, we extract the image corresponding to a certain field of view from the corrected image. By performing the process, we validate the improvement of detection distance. Furthermore, we attempt to detection pedestrians on the sidewalk around Waseda University. In the experiment, we adopt gnomonic projection as a projection method and YOLOv3 as a human detection.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) 360-degree video / Human detection / Congestion estimation
Paper # SRW2019-30,SeMI2019-74,CNR2019-24
Date of Issue 2019-10-29 (SRW, SeMI, CNR)

Conference Information
Committee SRW / SeMI / CNR
Conference Date 2019/11/5(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Kozo Keisaku Engineering Inc.
Topics (in Japanese) (See Japanese page)
Topics (in English) IoT Workshop (Oral or Poster presentation)
Chair Satoshi Denno(Okayama Univ.) / Susumu Ishihara(Shizuoka Univ.) / Kazunori Takashio(Keio Univ.)
Vice Chair Keiichi Mizutani(Kyoto Univ.) / Kazuya Monden(Hitachi) / Koji Yamamoto(Kyoto Univ.) / Masayuki Kanbara(NAIST) / Yoshihiko Murakawa(Fujitsu Labs.)
Secretary Keiichi Mizutani(Tokyo Inst. of Tech.) / Kazuya Monden(Anritsu) / Koji Yamamoto(Kyoto Univ.) / Masayuki Kanbara(NTT DOCOMO) / Yoshihiko Murakawa(Hitachi)
Assistant Masaaki Fuse(Anritsu) / Tomoki Murakami(NTT) / Akira Uchiyama(Osaka Univ.) / Kenji Kanai(Waseda Univ.) / Masafumi Hashimoto(Osaka Univ.) / Yuka Kobayashi(Toshiba) / Masanori Yokoyama(NTT)

Paper Information
Registration To Technical Committee on Short Range Wireless Communications / Technical Committee on Sensor Network and Mobile Intelligence / Technical Committee on Cloud Network Robotics
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) [Poster Presentation] Performance Evaluations of Human Detection Method using 360-degree Image toward Sidewalk Congestion Estimation
Sub Title (in English)
Keyword(1) 360-degree video
Keyword(2) Human detection
Keyword(3) Congestion estimation
1st Author's Name Kazuki Murayama
1st Author's Affiliation Waseda University(Waseda Univ)
2nd Author's Name Kenji Kanai
2nd Author's Affiliation Waseda University(Waseda Univ)
3rd Author's Name Masaru Takeuchi
3rd Author's Affiliation Waseda University(Waseda Univ)
4th Author's Name Jiro Katto
4th Author's Affiliation Waseda University(Waseda Univ)
Date 2019-11-06
Paper # SRW2019-30,SeMI2019-74,CNR2019-24
Volume (vol) vol.119
Number (no) SRW-265,SeMI-266,CNR-267
Page pp.pp.45-46(SRW), pp.65-66(SeMI), pp.43-44(CNR),
#Pages 2
Date of Issue 2019-10-29 (SRW, SeMI, CNR)