Yoshiki Kubotani
Yoshiki Kubotani has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Shigeo Morishima
Waseda Research Institute for Science and Engineering, Tokyo, Japan
4 shared publications - Frederick W. B. Li
Department of Computer Science University of Durham U.K
2 shared publications - Hina Otake
Waseda University
2 shared publications - Hubert P. H. Shum
Durham University, Durham, United Kingdom
2 shared publications - Tanqiu Qiao
Department of Computer Science Durham University DH1 3LE Durham UK
2 shared publications - Yoshihiro Fukuhara
Waseda University
2 shared publications
Latest Publications
- Open Access
- 3 June 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 290
Geometric visual fusion graph neural networks for multi-person human-object interaction recognition in videos
- 3 June 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 290
Open AccessGeometric visual fusion graph neural networks for multi-person human-object interaction recognition in videos
- 12 May 2025
- Published by Springer Nature
- p. 316-332
Detect Fake with Fake: Leveraging Synthetic Data-Driven Representation for Synthetic Image Detection
- 12 May 2025
- Published by Springer Nature
- p. 316-332
Detect Fake with Fake: Leveraging Synthetic Data-Driven Representation for Synthetic Image Detection
- Open Access
Detect Fake with Fake: Leveraging Synthetic Data-driven Representation for Synthetic Image Detection
Are general-purpose visual representations acquired solely from synthetic data useful for detecting fake images? In this work, we show the effectiveness of synthetic data-driven representations for synthetic image detection. Upon analysis, we find that vision transformers trained by the latest visual representation learners with synthetic data can effectively distinguish fake from real images without seeing any real images during pre-training. Notably, using SynCLR as the backbone in a state-of-the-art detection method demonstrates a performance improvement of +10.32 mAP and +4.73% accuracy over the widely used CLIP, when tested on previously unseen GAN models. Code is available at https://github.com/cvpaperchallenge/detect-fake-with-fake.
Open AccessDetect Fake with Fake: Leveraging Synthetic Data-driven Representation for Synthetic Image Detection
Are general-purpose visual representations acquired solely from synthetic data useful for detecting fake images? In this work, we show the effectiveness of synthetic data-driven representations for synthetic image detection. Upon analysis, we find that vision transformers trained by the latest visual representation learners with synthetic data can effectively distinguish fake from real images without seeing any real images during pre-training. Notably, using SynCLR as the backbone in a state-of-the-art detection method demonstrates a performance improvement of +10.32 mAP and +4.73% accuracy over the widely used CLIP, when tested on previously unseen GAN models. Code is available at https://github.com/cvpaperchallenge/detect-fake-with-fake.
- 28 October 2022
- Published by Springer Nature
- p. 474-491
Geometric Features Informed Multi-person Human-Object Interaction Recognition in Videos
- 28 October 2022
- Published by Springer Nature
- p. 474-491
Geometric Features Informed Multi-person Human-Object Interaction Recognition in Videos