Xingyu Wan

Title: Innovations of Xingyu Wan in Multi-Object Tracking

Introduction

Xingyu Wan is an accomplished inventor based in Xi'an, China. He has made significant contributions to the field of multi-object tracking through his innovative patent. His work focuses on enhancing tracking accuracy using advanced machine learning techniques.

Latest Patents

Xingyu Wan holds a patent for a framework titled "Multi-object tracking using online metric learning with long short-term memory." This framework employs a dual Long Short-Term Memory (LSTM) network designed for multi-object tracking (MOT). The dual LSTM network learns metrics alongside an online updating scheme for data association. It fuses relevant features of trajectories to interpret both temporal and spatial components non-linearly, concurrently outputting a similarity score. Each trajectory can be initialized with an LSTM model, and the metric is updated in real-time during the tracking phase. The framework utilizes Optical Flow and affine transformations to generate robust tracklets for initialization, resulting in increased tracking accuracy. Furthermore, the MOT framework demonstrates improved performance and flexibility for use in various scenarios.

Career Highlights

Xingyu Wan is currently associated with Deep North, Inc., where he continues to develop innovative solutions in the field of visual tracking. His expertise in machine learning and tracking technologies has positioned him as a key player in his organization.

Collaborations

Xingyu collaborates with Jinjun Wang, contributing to advancements in their shared field of expertise.

Conclusion

Xingyu Wan's innovative work in multi-object tracking showcases his commitment to enhancing tracking technologies through advanced machine learning methods. His contributions are paving the way for more accurate and efficient tracking solutions in various applications.