Mantra's Ryosuke Furuta won the Excellence Award and Audience Award at the 29th Meeting on Image Recognition and Understanding (MIRU2026)!
At the 29th Meeting on Image Recognition and Understanding (MIRU2026), held at Dejima Messe Nagasaki from August 3 to August 6, 2026, Mantra's Ryosuke Furuta received both the MIRU Excellence Award and the MIRU Audience Award!
©Kato Masaki ©Yabuno Tenya, Watanabe Tatsuya
MIRU Excellence Award
OS1C-06: Inference-time Flow Trajectory Optimization for Manga Image Editing Ryosuke Furuta (Mantra, The University of Tokyo)
Reason for Award: Proposed a clever and straightforward technique to adapt a Rectified Flow model pre-trained on natural images for manga image editing across domains without re-training. The problem formulation thoughtfully accounts for copyright concerns unique to manga, and the concept of enabling training-free, out-of-domain image editing was highly recognized for its significant industrial and academic value, with promising applicability to non-manga domains.
MIRU Audience Award
OS1C-06: Inference-time Flow Trajectory Optimization for Manga Image Editing Ryosuke Furuta (Mantra, The University of Tokyo)
[Click here for award ceremony details]
About the Meeting on Image Recognition and Understanding (MIRU)
The Meeting on Image Recognition and Understanding (MIRU) is Japan's largest academic conference dedicated to image recognition and understanding technologies. It serves as a forum where researchers, engineers, and students from various companies and universities gather to discuss, network, and present the latest research spanning from fundamentals to practical applications.
- Paper Title:
Inference-time Trajectory Optimization for Structure-Preserving Manga Image Editing - Paper: https://arxiv.org/abs/2603.27790
- Code: https://github.com/mantra-inc/structure-preserving-manga-editing