Papers
14
Total Citations
173
H-Index
8
About
Fuyuki Tokuda is a robotics researcher whose work sits at the intersection of computer vision, deep learning, and robotic manipulation. His research spans three principal areas: visual servoing, 3D perception for robot grasping, and textile automation — a combination that positions him as a distinctive voice in intelligent robotic systems. Tokuda's most influential contribution is DEFINet, a convolutional neural network architecture for visual servoing that computes robot control inputs directly from images, eliminating the need for hand-crafted feature engineering. This work has garnered 46 citations and spawned a productive line of follow-up research. Complementing this, his studies on 3D instance segmentation and 6D pose estimation — including occlusion-aware bin-picking algorithms — address core challenges in industrial robotics, collectively accumulating nearly 60 additional citations. Perhaps most distinctively, Tokuda has turned his attention to garment manufacturing automation, developing robotic sewing systems with time-scaling control, fixture-free dual-arm sewing architectures, fabric-folding end-effectors, and a passive actuator-less gripper for handling delicate fabric parts. This body of work, totaling over 140 citations across his career, demonstrates a rare ability to bridge foundational perception research with highly practical, real-world automation challenges in the apparel industry.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Time-Scaling Modeling and Control of Robotic Sewing System11 citations · 2024
- 6Object Positioning by Visual Servoing Based on Deep Learning11 citations · 2019
- 7Fixture-Free 2D Sewing Using a Dual-Arm Manipulator System10 citations · 2024
- 8Robot End-effector for Fabric Folding9 citations · 2023
- 9Neural Network based Visual Servoing for Eye-to-Hand Manipulator8 citations · 2020
- 10Passive Actuator-Less Gripper for Pick-and-Place of a Piece of Fabric7 citations · 2025