| Presentation | 2019-12-06 Adversarial Examples for Monocular Depth Estimation Koichiro Yamanaka, Ryutaroh Matsumoto, Keita Takahashi, Toshiaki Fujii, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Adversarial examples for classification and object recognition problems using convolutional neural net- works (CNN) have attracted much attention in recent years. By adding perturbations to an input image of a CNN, adversarial attack is able to intentionally induce erroneous inferences. Adversarial attack is roughly classified into two types. The one is a method that slightly changes the pixel values of an entire input image, and the other is a method that overwrites a specific pattern (adversarial patch) on a local region of the input image. The latter method is called the adversarial patch attack, and recently real world attack was proposed by using this method. In other words, the classification CNN and the object recognition CNN could be deceived by taking a printed adversarial patch with a camera. However, adversarial attacks on regression problems have not been studied well. In this paper, we propose an adversarial attack method for regression problem, especially for monocular depth estimation CNN. We demonstrate that our method is capable of generating adversarial patches that can arbitrarily manipulate the output of the monocular depth estimation. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Monocular Depth Estimation / CNN / Adversarial Examples / Adversarial Patch |
| Paper # | CS2019-83,IE2019-63 |
| Date of Issue | 2019-11-28 (CS, IE) |
| Conference Information | |
| Committee | IE / CS / IPSJ-AVM / ITE-BCT |
|---|---|
| Conference Date | 2019/12/5(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Aiina Center |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Image coding, Communications and streaming technologies, etc. |
| Chair | Hideaki Kimata(NTT) / Hidenori Nakazato(Waseda Univ.) / Sei Naito(KDDI Research, Inc.) / Kyoichi Saito(NHK) |
| Vice Chair | Kazuya Kodama(NII) / Keita Takahashi(Nagoya Univ.) / Jun Terada(NTT) / / Eichi Murata(Kyoto Univ.) / Tomomi Fukazawa(TBS) |
| Secretary | Kazuya Kodama(NTT) / Keita Takahashi(NHK) / Jun Terada(Waseda Univ.) / (Mitsubishi Electric) / Eichi Murata(NTT) / Tomomi Fukazawa(Tokyo Univ. of Science) |
| Assistant | Kyohei Unno(KDDI Research) / Norishige Fukushima(Nagoya Inst. of Tech.) / Kazutaka Hara(NTT) / Hiroyuki Saito(OKI) / / Minoru Okada(Nara Institute of Science and Technology) / Jun Yukawa(Mitsubishi Electric Corporation) / Takao Tsuda(NHK) |
| Paper Information | |
| Registration To | Technical Committee on Image Engineering / Technical Committee on Communication Systems / Special Interest Group on Audio Visual and Multimedia Information Processing / Technical Group on Broadcasting and Communication Technologies |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Adversarial Examples for Monocular Depth Estimation |
| Sub Title (in English) | |
| Keyword(1) | Monocular Depth Estimation |
| Keyword(2) | CNN |
| Keyword(3) | Adversarial Examples |
| Keyword(4) | Adversarial Patch |
| 1st Author's Name | Koichiro Yamanaka |
| 1st Author's Affiliation | Nagoya University(Nagoya Univ.) |
| 2nd Author's Name | Ryutaroh Matsumoto |
| 2nd Author's Affiliation | Nagoya University(Nagoya Univ.) |
| 3rd Author's Name | Keita Takahashi |
| 3rd Author's Affiliation | Nagoya University(Nagoya Univ.) |
| 4th Author's Name | Toshiaki Fujii |
| 4th Author's Affiliation | Nagoya University(Nagoya Univ.) |
| Date | 2019-12-06 |
| Paper # | CS2019-83,IE2019-63 |
| Volume (vol) | vol.119 |
| Number (no) | CS-323,IE-324 |
| Page | pp.pp.91-95(CS), pp.91-95(IE), |
| #Pages | 5 |
| Date of Issue | 2019-11-28 (CS, IE) |