Jianyu Yang

About

Jianyu Yang is a researcher whose work sits at the intersection of computer vision, robotics, and human-robot interaction, with particular expertise in motion recognition, object recognition, and embedded perception systems. His research spans over a decade, addressing fundamental challenges in how machines perceive and interpret the physical world. Yang's early contributions focused on developing robust descriptors for 3D motion trajectory recognition, most notably the Mixed Signature — an invariant descriptor that captures rich motion information from human gestures and robotic actions, earning 17 citations and establishing a foundation for his trajectory-based recognition framework. This line of work evolved into segmentation and description methods for real-time human motion recognition, directly supporting advances in human-robot interaction. In parallel, Yang made significant strides in object recognition, exploring metric learning approaches, salient shape contour detection, and hybrid shape descriptors tailored for robot vision applications. His investigation into sparse representation methods using constrained Restricted Boltzmann Machines contributed meaningfully to cognitive robotic perception. His most cited work, FastHand (27 citations), demonstrates his ability to translate complex perception algorithms into practical, efficient solutions deployable on embedded systems — a critical bridge between research and real-world robotics. Collectively, Yang's portfolio reflects a consistent commitment to making machine perception faster, smarter, and more applicable across diverse robotic environments.

Research Focus

Key Achievements

Top Papers

1. 1
   FastHand: Fast monocular hand pose estimation on embedded systems

   27 citations · 2021
2. 2
   Mixed Signature: An Invariant Descriptor for 3D Motion Trajectory Perception and Recognition

   17 citations · 2011
3. 3
   Metric learning based object recognition and retrieval

   16 citations · 2016
4. 4
   Analysis of Different Sparsity Methods in Constrained RBM for Sparse Representation in Cognitive Robotic Perception

   15 citations · 2015
5. 5
   A new descriptor for 3D trajectory recognition via modified CDTW

   8 citations · 2010
6. 6
   Finding salient points of shape contour for object recognition

   6 citations · 2015
7. 7
   A hybrid shape descriptor for object recognition

   4 citations · 2015
8. 8
   On trajectory segmentation and description for motion recognition

   4 citations · 2015
9. 9
   An obstacle avoidance method based on non-radial arrangement of distance sensors for vacuum cleaning robot

   4 citations · 2016
10. 10
    Vision Based Hand Gesture Recognition Using 3D Shape Context

    3 citations · 2018

Key Collaborators

Contact & Links