Presentation | 2017-10-05 Automatic Segmentation using Graph Cuts based on Human Body Detection of Kinect V2 Kodai Kobayashi, Masatoshi Sato, Takuya Asaka, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | To quickly segment human body region in images or videos with high accuracy makes video surveillances and making composite images efficient. In conventional method, to segment region without location information given by users is difficult. In this paper, we present a novel method to automatically segment human body region by using Graph Cuts based on human body detection of Kinect V2. By applying expansion and contraction processing on the image of human body detection and automating labeling "seed", we can automate segmentation by Graph Cuts. The conventional methods of automatic segmentation have a problem of slow processing time. The proposed method can accelerate processing time by replacing Graph Cuts with Maximum-Flow Neural Network(MF-NN). When MF-NN is designed with the integrated circuit such as Programmable Logic Device, Memristor or Phase Change Memory which can change graph structure and branch conductance, the novel min-cut solution with real– time processing can be expected. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Graph Cuts / segmentation / Kinect V2 |
Paper # | IE2017-50 |
Date of Issue | 2017-09-28 (IE) |
Conference Information | |
Committee | IE / ITE-ME / ITE-AIT |
---|---|
Conference Date | 2017/10/5(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Takayuki Hamamoto(Tokyo Univ. of Science) / Miki Haseyama(Hokkaido Univ.) / Nobuhiko Mukai(Tokyo Cisy Univ.) |
Vice Chair | Kazuya Kodama(NII) / Hideaki Kimata(NTT) |
Secretary | Kazuya Kodama(Nagoya Univ.) / Hideaki Kimata(KDDI Research) |
Assistant | Yasutaka Matsuo(NHK) / Kazuya Hayase(NTT) |
Paper Information | |
Registration To | Technical Committee on Image Engineering / Technical Group on Media Engineering / Technical Group on Artistic Image Technology |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Automatic Segmentation using Graph Cuts based on Human Body Detection of Kinect V2 |
Sub Title (in English) | |
Keyword(1) | Graph Cuts |
Keyword(2) | segmentation |
Keyword(3) | Kinect V2 |
1st Author's Name | Kodai Kobayashi |
1st Author's Affiliation | Tokyo Metropolitan University(Tokyo Metropolitan Univ) |
2nd Author's Name | Masatoshi Sato |
2nd Author's Affiliation | Tamagawa University(Tamagawa Univ) |
3rd Author's Name | Takuya Asaka |
3rd Author's Affiliation | Tokyo Metropolitan University(Tokyo Metropolitan Univ) |
Date | 2017-10-05 |
Paper # | IE2017-50 |
Volume (vol) | vol.117 |
Number (no) | IE-228 |
Page | pp.pp.19-24(IE), |
#Pages | 6 |
Date of Issue | 2017-09-28 (IE) |