| Presentation | 2022-03-07 Extention of robust image classification system with Adversarial Example Detectors Miki Tanaka, Takayuki Osakabe, Hitoshi Kiya, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In image classification with deep learning, there is a risk that an attacker can intentionally manipulate the prediction results of image classification by using images with a small designed noise, called adversarial examples. In this paper, we propose a robust image classification system against adversarial examples. In order to prevent the attack, there are two approaches: using a robust classifier, and using a detection method of adversarial examples. A robust image classification system with the combination of these two approaches was demonstrated to outperform conventional methods. In this paper, we extend the robust image classification system with the combination of the two approaches by using multiple features. The proposed method is more robust against various adversarial attacks and noise levels than the conventional one. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Adversarial example / Machine learning / Deep learning / Adversarial detection |
| Paper # | EMM2021-105 |
| Date of Issue | 2022-02-28 (EMM) |
| Conference Information | |
| Committee | EMM |
|---|---|
| Conference Date | 2022/3/7(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | (Primary: Online, Secondary: On-site) |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Image and Sound Quality, Metrics for Perception and Recognition, Human Auditory and Visual System, etc. |
| Chair | Ryoichi Nishimura(NICT) |
| Vice Chair | Masaaki Fujiyoshi(Tokyo Metropolitan Univ.) / Masatsugu Ichino(Univ. of Electro-Comm.) |
| Secretary | Masaaki Fujiyoshi(Utsunomiya Univ.) / Masatsugu Ichino(NICT) |
| Assistant | Shoko Imaizumi(Chiba Univ.) / Youichi Takashima(Kaishi Professional Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Enriched MultiMedia |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Extention of robust image classification system with Adversarial Example Detectors |
| Sub Title (in English) | |
| Keyword(1) | Adversarial example |
| Keyword(2) | Machine learning |
| Keyword(3) | Deep learning |
| Keyword(4) | Adversarial detection |
| 1st Author's Name | Miki Tanaka |
| 1st Author's Affiliation | Tokyo Metropolitan University(Tokyo Metro. Univ.) |
| 2nd Author's Name | Takayuki Osakabe |
| 2nd Author's Affiliation | Tokyo Metropolitan University(Tokyo Metro. Univ.) |
| 3rd Author's Name | Hitoshi Kiya |
| 3rd Author's Affiliation | Tokyo Metropolitan University(Tokyo Metro. Univ.) |
| Date | 2022-03-07 |
| Paper # | EMM2021-105 |
| Volume (vol) | vol.121 |
| Number (no) | EMM-417 |
| Page | pp.pp.76-80(EMM), |
| #Pages | 5 |
| Date of Issue | 2022-02-28 (EMM) |