Papers
22
Total Citations
435
H-Index
10
About
Kentaro Wada is a robotics and computer vision researcher whose work sits at the intersection of 3D scene understanding, robotic manipulation, and deep learning. His research primarily addresses the fundamental challenges of enabling robots to perceive, grasp, and manipulate objects in complex, cluttered environments — problems central to warehouse automation and general-purpose robotics. Wada's most influential contribution, *MoreFusion* (2020, 103 citations), advanced the field of 6D object pose estimation by combining multi-object reasoning with volumetric fusion, allowing robots to better understand scenes with occlusions. His work on *Coarse-to-Fine Q-attention* (2022, 66 citations) demonstrated a more stable and data-efficient approach to learning robotic manipulation through discretised reinforcement learning. His hardware contributions are equally notable — his suction-and-finger hybrid gripper designs (2017, 51 citations) tackled the real-world challenge of grasping diverse objects in tight spaces, directly addressing industrial automation needs. Across his portfolio, Wada has consistently bridged perception and action, developing systems for instance segmentation under occlusion, grasp modality fusion, and object reorientation for precise placement. With over 350 cumulative citations, his work has meaningfully shaped how modern robotic systems see and interact with their physical environments.
Research Focus
Key Achievements
Top Papers
- 1MoreFusion: Multi-object Reasoning for 6D Pose Estimation from Volumetric Fusion103 citations · 2020
- 2
- 3
- 4
- 5ReorientBot: Learning Object Reorientation for Specific-Posed Placement24 citations · 2022
- 6
- 7
- 8
- 9
- 10