| Presentation | 2023-03-02 An image watermarking method using adversarial perturbations Sei Takano, Mitsuji Muneyasu, Soh Yoshida, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The performance of convolutional neural networks (CNNs) has been dramatically improved in recent years, and they have attracted much attention in object detection, image classification, and others. On the other hand, it is known that there exist perturbations, called adversarial perturbations, which cause CNNs to misidentify objects and are almost indistinguishable from humans. If perturbations are considered watermarking information, they can be used to add some information to an image. Based on this idea, we propose a new watermarking method that assumes a CNN as a detection method. Considering compression resistance, we develop a C&W-DCT attack that perturbs the DCT coefficients and propose a new watermarking method. Through experiments, we show that the proposed method can be used as a watermark technique in practice and is resistant to JPEG compression. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | adversarial perturbation / deep learning / adversarial example / target attack / MobileNetV2 / JPEG |
| Paper # | SIS2022-43 |
| Date of Issue | 2023-02-23 (SIS) |
| Conference Information | |
| Committee | SIS |
|---|---|
| Conference Date | 2023/3/2(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Chiba Institute of Technology |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Tomoaki Kimura(Kanagawa Inst. of Tech.) |
| Vice Chair | Naoto Sasaoka(Tottori Univ.) / Hakaru Tamukoh(Kyushu Inst. of Tech.) |
| Secretary | Naoto Sasaoka(NTT) / Hakaru Tamukoh(Kansai Univ.) |
| Assistant | Yoshiaki Makabe(Kanagawa Inst. of Tech.) / Yosuke Sugiura(Saitama Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Smart Info-Media Systems |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | An image watermarking method using adversarial perturbations |
| Sub Title (in English) | |
| Keyword(1) | adversarial perturbation |
| Keyword(2) | deep learning |
| Keyword(3) | adversarial example |
| Keyword(4) | target attack |
| Keyword(5) | MobileNetV2 |
| Keyword(6) | JPEG |
| 1st Author's Name | Sei Takano |
| 1st Author's Affiliation | Kansai University(Kansai Univ.) |
| 2nd Author's Name | Mitsuji Muneyasu |
| 2nd Author's Affiliation | Kansai University(Kansai Univ.) |
| 3rd Author's Name | Soh Yoshida |
| 3rd Author's Affiliation | Kansai University(Kansai Univ.) |
| Date | 2023-03-02 |
| Paper # | SIS2022-43 |
| Volume (vol) | vol.122 |
| Number (no) | SIS-410 |
| Page | pp.pp.15-20(SIS), |
| #Pages | 6 |
| Date of Issue | 2023-02-23 (SIS) |