Presentation | 2021-03-02 A Face Identification System Based on Self-Supervised Learning Using Triplet-based Variational Autoencoder Yuta Hagio, Yutaka Kaneko, Yuta Hoshi, Marina Kamimura, Yasuhiro Murasaki, Masao Yamatomo, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In this paper, we propose a face image identification system based on self-supervised learning for IoT devices and companion robots used in the home. Our system automatically constructs a triplet-based training dataset, which consists of two face images of the same person and one face image of a different person. In addition, a face image discriminator based on Triplet-based Variational Autoencoder is trained using the constructed dataset. Experimental results show that our system could automatically construct a training data, and about 98% of the constructed data was in the correct format, consisting of two face images of the same person and one face image of a different person. We also confirmed that our system could identify four face images with 94.2% accuracy and six face images with 95.2% accuracy, by training the constructed dataset. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Face Identification / Self-Supervised Learning / Variational Autoencoder / Internet of Things / Companion Robots |
Paper # | BioX2020-46,CNR2020-19 |
Date of Issue | 2021-02-23 (BioX, CNR) |
Conference Information | |
Committee | BioX / CNR |
---|---|
Conference Date | 2021/3/2(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Online |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Akira Otsuka(AIST) / Kazunori Takashio(Keio Univ.) |
Vice Chair | Takahiro Aoki(Fujitsu Labs.) / Masatsugu Ichino(Univ. of Electro-Comm.) / Masayuki Kanbara(NAIST) / Yoshihiko Murakawa(Fujitsu Labs.) |
Secretary | Takahiro Aoki(SECOM) / Masatsugu Ichino(KDDI Research) / Masayuki Kanbara(Shibaura Inst. of Tech.) / Yoshihiko Murakawa(Panasonic) |
Assistant | Emiko Sano(MitsubishiElectric) / Akihiro Hayasaka(NEC) / Yuka Kobayashi(Toshiba) / Masanori Yokoyama(NTT) |
Paper Information | |
Registration To | Technical Committee on Biometrics / Technical Committee on Cloud Network Robotics |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | A Face Identification System Based on Self-Supervised Learning Using Triplet-based Variational Autoencoder |
Sub Title (in English) | |
Keyword(1) | Face Identification |
Keyword(2) | Self-Supervised Learning |
Keyword(3) | Variational Autoencoder |
Keyword(4) | Internet of Things |
Keyword(5) | Companion Robots |
1st Author's Name | Yuta Hagio |
1st Author's Affiliation | Japan Broadcasting Corporation(NHK) |
2nd Author's Name | Yutaka Kaneko |
2nd Author's Affiliation | Japan Broadcasting Corporation(NHK) |
3rd Author's Name | Yuta Hoshi |
3rd Author's Affiliation | Japan Broadcasting Corporation(NHK) |
4th Author's Name | Marina Kamimura |
4th Author's Affiliation | Japan Broadcasting Corporation(NHK) |
5th Author's Name | Yasuhiro Murasaki |
5th Author's Affiliation | Japan Broadcasting Corporation(NHK) |
6th Author's Name | Masao Yamatomo |
6th Author's Affiliation | Japan Broadcasting Corporation(NHK) |
Date | 2021-03-02 |
Paper # | BioX2020-46,CNR2020-19 |
Volume (vol) | vol.120 |
Number (no) | BioX-393,CNR-394 |
Page | pp.pp.32-37(BioX), pp.32-37(CNR), |
#Pages | 6 |
Date of Issue | 2021-02-23 (BioX, CNR) |