Cheng Lu

Title: Innovations by Cheng Lu in Human Motion Prediction and Feature Matching

Introduction

Cheng Lu is a prominent inventor based in Guangdong, China. He has made significant contributions to the fields of human motion prediction and feature matching through his innovative patents. With a total of three patents to his name, Cheng Lu is recognized for his advanced technological solutions.

Latest Patents

Cheng Lu's latest patents include a "System and method for performing unified human motion prediction (HMP) for 2D and 3D poses." This system provides a unified HMP module that encodes multiple existing two-dimensional (2D) and three-dimensional (3D) pose sequences into a shared latent space. The module also includes a decoder that predicts future 2D and 3D pose sequences based on the encoded data. The training process is conducted in a weakly supervised manner, minimizing the loss between predicted and actual future poses.

Another notable patent is the "System for three-dimensional geometric guided student-teacher feature matching (3DG-STFM)." This system features a multi-modal teacher model that determines feature matching between pairs of RGB-D images. It also includes a mono-modal student model that learns from the teacher model to enhance feature matching accuracy. The system employs both coarse and fine-level knowledge transfer loss functions to guide the learning process effectively.

Career Highlights

Cheng Lu is currently employed at Guangzhou Xiaopeng Autopilot Technology Co., Ltd. His work at this company focuses on developing cutting-edge technologies that enhance autonomous driving systems. His innovative approaches have positioned him as a key player in the field of artificial intelligence and machine learning.

Collaborations

Cheng Lu collaborates with talented individuals such as Chen Bai and Chengzhang Zhong. These partnerships foster a creative environment that drives innovation and technological advancement.

Conclusion

Cheng Lu's contributions to the fields of human motion prediction and feature matching demonstrate his expertise and commitment to innovation. His patents reflect a deep understanding of complex systems and a dedication to advancing technology.