Seitaro Otsuki
Papers
2
Total Citations
7
H-Index
2
About
Seitaro Otsuki is a robotics researcher advancing human-robot interaction through multimodal perception and language understanding. His work focuses on enabling robots to safely and intuitively collaborate with humans, particularly in domestic service and handover scenarios. Otsuki’s key contributions include developing a shared Transformer encoder with mask-based 3D model estimation for container mass estimation (2022, 5 citations), which allows robots to safely handle objects during human-to-robot handovers by accurately inferring physical properties like filling level and type. He also pioneered prototypical contrastive transfer learning for multimodal language understanding (2023, 2 citations), addressing the challenge of robots interpreting ambiguous natural language commands—such as “Bring me a bottle from the kitchen”—by grounding language in visual and physical context. This work bridges the gap between high-level instructions and robotic action, a critical step toward assistive robots that can operate in unstructured home environments. Otsuki’s research sits at the intersection of computer vision, natural language processing, and safe robot control, with implications for elderly care and household automation. His transformer-based approaches demonstrate how deep learning can make robots more perceptive and responsive to human needs.
Research Focus
Key Achievements
Top Papers
- 1
- 2