Scholar

Okatani Takayuki

Google Scholar ID: gn780jcAAAAJ

Professor, Graduate School of Information Sciences, Tohoku University / RIKEN Center for AIP

Computer vision

Citations & Impact

All-time

Citations

3,945

H-index

i10-index

Publications

Co-authors

list available

Contact

No contact links provided.

Publications

20 items

Browse publications on Google Scholar (top-right) ↗

Resume (English only)

Academic Achievements

Contributions to the development of object detection, multimodal AI, video super-resolution, semantic segmentation, SLAM, visual localization, anomaly detection, image captioning, lightweight image recognition models, bridge inspection systems, texture anomaly detection, and driving hazard prediction.

Research Experience

Self-supervised learning for object detection with content and location integration; Who decides what to detect? Challenges and new problem settings in OSOD research; Landslide image analysis and disaster risk assessment using multimodal AI; RefVSR++: A high-precision video super-resolution method utilizing reference videos; Domain adaptation for semantic segmentation; Fast bundle adjustment for vehicle SLAM using graph networks; Fast and accurate visual localization through local feature integration; Embodied visual navigation leveraging problem symmetry; Logical and structural anomaly detection from images based on local and global knowledge integration; GRIT: Grid + Region feature-based image captioning; SBCFormer: Lightweight image recognition model operable on single-board computers; Interactive multimodal AI for bridge inspection; Zero-shot texture anomaly detection; Driving hazard prediction based on traffic scene understanding.

Miscellany