Atsushi Sagata
Papers
1
Total Citations
2
H-Index
1
About
Atsushi Sagata has made significant contributions to robot vision technology, focusing on object instance recognition and 3D pose estimation—critical components for enabling robots to perceive and interact with their environments. His most-cited work, "Adaptive Loss Balancing for Multitask Learning of Object Instance Recognition and 3D Pose Estimation" (2019), addresses a key challenge in multitask learning: the integration of loss functions for simultaneous tasks. Sagata proposed an adaptive loss balancing method that dynamically adjusts weighting parameters during training, improving the accuracy of both instance recognition and pose estimation over unified balancing approaches. This innovation enhances the robustness of robot vision systems in real-world applications. With 2 citations, his work is foundational for researchers exploring multitask learning in robotics. Sagata’s research bridges the gap between theoretical optimization and practical deployment, offering a framework that adapts to task-specific demands. His achievements underscore a commitment to advancing autonomous systems, making his contributions valuable for students and engineers developing next-generation robotic perception technologies.
Research Focus
Key Achievements
Top Papers
- 1