Tohru Kamiya
Papers
6
Total Citations
92
H-Index
4
About
Tohru Kamiya is a leading researcher at the intersection of computer vision and robotics, whose work primarily focuses on semantic segmentation, deformable object manipulation, and robotic navigation. His most impactful contribution, "WideSegNeXt" (2020, 58 citations), advanced semantic image segmentation by integrating wide residual networks with NeXt dilated units, significantly improving performance for applications in autonomous driving, robotic picking, and medical imaging. Kamiya has also made notable strides in robotic control, developing a characteristics-based visual servo system for 6DOF robot arms (2021, 11 citations) that enhances precision in automated tasks. His research extends to 3D object detection using improved PointRCNN (2022) and underwater image super-resolution via SRCNN (2021), demonstrating versatility across challenging environments. Most recently, Kamiya addressed multi-robot navigation in dynamic crowded settings (2024), proposing a velocity prediction algorithm that combines BP neural networks with reciprocal velocity obstacles to enable conflict-free path planning. With a growing body of work that bridges deep learning and practical robotics, Kamiya's contributions are shaping how machines perceive and interact with complex, real-world scenes.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Characteristics based visual servo for 6DOF robot arm control11 citations · 2021
- 43D object detection using improved PointRCNN4 citations · 2022
- 5Underwater image super-resolution using SRCNN3 citations · 2021
- 6