Kohei Fujita

Papers

Total Citations

H-Index

About

Kohei Fujita is a researcher advancing the field of robotic perception and computer vision, with a primary focus on object pose estimation for automated manipulation. His most cited work, "PYNet: Poseclass and Yaw Angle Output Network for Object Pose Estimation" (2023, 10 citations), tackles a critical challenge in robotics: enabling robots to accurately grasp simple-shaped objects, such as retail goods, using RGBD cameras. Conventional methods often struggle with estimating the three-dimensional poses of such objects due to their lack of distinctive features. Fujita’s contribution lies in developing a deep learning framework that efficiently outputs both pose class and yaw angle, simplifying the estimation process for real-world robotic applications. This work directly addresses the practical needs of warehouse automation and retail logistics, where reliable object grasping is essential. While his citation count is still growing, the relevance of his research to the booming field of service robotics and e-commerce automation underscores its potential impact. Fujita’s focus on bridging the gap between algorithmic innovation and industrial deployment marks him as a promising contributor to intelligent robotic systems.