| Presentation | 2020-10-09 Style Adapted DataBase: Generalizing Hand Segmentation via Semantics-aware Stylization Takehiko Ohkawa, Takuma Yagi, Yoichi Sato, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Domain shift in first-person vision degrades the performance of hand segmentation, which is caused by changes in lighting conditions and background appearances. In this paper, we propose a semantics-aware stylization approach for domain adaptation using an image style transfer with only a few target labels. Specifically, given a source image as content and a target image as style, foreground and background are separated by their labels, and the network transfers the styles of the target image to the source image separately for the foreground and background. Multiple styles can be fed into a stylized source dataset, thus the model trained on the dataset simultaneously generalizes to multiple target domains at once. Our method achieves the best cross-dataset generalization against the state-of-the-art domain adaptation methods for hand segmentation. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Domain Adaptation / Style Transfer / Hand Segmentation / First Person Vision |
| Paper # | PRMU2020-23 |
| Date of Issue | 2020-10-02 (PRMU) |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2020/10/9(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Virtual |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Recognition and understating of human |
| Chair | Yoichi Sato(Univ. of Tokyo) |
| Vice Chair | Akisato Kimura(NTT) / Masakazu Iwamura(Osaka Pref. Univ.) |
| Secretary | Akisato Kimura(Mobility Technologies) / Masakazu Iwamura(Chubu Univ.) |
| Assistant | Takashi Shibata(NTT) / Masashi Nishiyama(Tottori Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Pattern Recognition and Media Understanding |
|---|---|
| Language | ENG-JTITLE |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Style Adapted DataBase: Generalizing Hand Segmentation via Semantics-aware Stylization |
| Sub Title (in English) | |
| Keyword(1) | Domain Adaptation |
| Keyword(2) | Style Transfer |
| Keyword(3) | Hand Segmentation |
| Keyword(4) | First Person Vision |
| 1st Author's Name | Takehiko Ohkawa |
| 1st Author's Affiliation | The University of Tokyo(UTokyo) |
| 2nd Author's Name | Takuma Yagi |
| 2nd Author's Affiliation | The University of Tokyo(UTokyo) |
| 3rd Author's Name | Yoichi Sato |
| 3rd Author's Affiliation | The University of Tokyo(UTokyo) |
| Date | 2020-10-09 |
| Paper # | PRMU2020-23 |
| Volume (vol) | vol.120 |
| Number (no) | PRMU-187 |
| Page | pp.pp.26-31(PRMU), |
| #Pages | 6 |
| Date of Issue | 2020-10-02 (PRMU) |