Max Robotics — Global Service Robot Directory + US Certification

About

Driton Salihu is a rising researcher at the intersection of computer vision, robotics, and 3D perception, whose work focuses on enabling machines to understand and interact with complex, dynamic environments. His key research areas include action detection and segmentation, 3D point cloud registration, and multimodal perception for robotics. Salihu’s major contributions lie in developing novel deep learning architectures that model spatiotemporal relationships and hierarchical structures. Notably, his work on "Modeling Action Spatiotemporal Relationships Using Graph-Based Class-Level Attention Network" (2023, 9 citations) advances long-term action detection by capturing dependencies between action classes, a critical capability for human-robot collaboration. He has also pioneered innovative approaches like Timestamp Supervised Contrastive Learning for action segmentation (2024) and Hierarchical Equivariant Graph Neural Networks for 9DoF point cloud registration (2024), both of which address fundamental challenges in robotic perception. His recent work extends to novel sensor modalities, including mmWave radar for surface material classification (2025) and hyperbolic contrastive learning for CAD model retrieval (2025). With a growing citation impact and a focus on practical robotic applications—from assistive robots to digital twinning—Salihu is establishing himself as a versatile and forward-thinking contributor to embodied AI and autonomous systems.