Zheng Liu

Title: Innovations by Zheng Liu in Video Analytics

Introduction

Zheng Liu is a prominent inventor based in Los Altos, CA, known for his contributions to the field of video analytics. With a total of 5 patents to his name, Liu has made significant advancements in object tracking technologies that enhance the capabilities of intelligent video systems.

Latest Patents

One of Liu's latest patents is focused on "Multi-object tracking using correlation filters in video analytics applications." This innovation involves learning correlation filters for object tracking based on occlusion maps. By applying an occlusion map to the image area, Liu's method effectively masks, excludes, and blurs occlusions of the target object. This approach allows for the learning of the target object from exposed portions while minimizing the impact of occlusions.

Another notable patent is "Real-time persistent object tracking for intelligent video analytics systems." This invention describes apparatuses and techniques for tracking the state of an object in an environment using a series of images generated over time. The system predicts future states of the object and updates identifiers based on the detection of other objects in the environment, ensuring accurate tracking even when the object is temporarily undetectable.

Career Highlights

Zheng Liu is currently employed at Nvidia Corporation, a leading company in graphics processing and AI technologies. His work at Nvidia has allowed him to apply his innovative ideas in a practical setting, contributing to the development of advanced video analytics systems.

Collaborations

Liu collaborates with talented individuals such as Joonhwa Shin and Kaustubh Purandare, who share his passion for pushing the boundaries of technology in video analytics.

Conclusion

Zheng Liu's contributions to the field of video analytics through his innovative patents demonstrate his expertise and commitment to advancing technology. His work continues to influence the development of intelligent systems that enhance our ability to analyze and interpret video data effectively.