Cheng-Yang Fu

University of North Carolina at Chapel Hill

Papers

Total Citations

H-Index

About

Cheng-Yang Fu is a leading researcher in computer vision, with a focus on object detection, 3D pose estimation, and instance-level recognition. His work bridges the gap between general object detection and practical, application-driven vision systems. In his highly cited 2016 paper, "Fast Single Shot Detection and Pose Estimation" (25 citations), Fu pioneered a convolutional network architecture that simultaneously detects objects and estimates their 3D pose in a single forward pass—a critical advance for navigation and robotics. This work demonstrated that pose estimation could be integrated directly into detection pipelines without sacrificing speed. Building on this, his 2018 paper "Target Driven Instance Detection" (17 citations) addressed a key challenge in household robotics: recognizing specific object instances rather than generic categories. Fu’s approach was designed to efficiently identify a few target objects in cluttered environments, making it highly practical for real-world deployment. His contributions have directly influenced the development of more specialized, efficient vision systems that move beyond broad detection to solve targeted, instance-level problems. Fu’s research continues to shape how machines perceive and interact with their environments, with lasting impact on both autonomous systems and robotics.