Continuous Pose for Monocular Cameras in Neural Implicit Representation
Modelling continuous pose with neural network and joint training with NeRF applications.
“Was there anything more exciting in life than seeking answers?” — Isaac Asimov.
I am a master student in robotics at ETH Zurich. I am currently a core member at AMZ driverless team and a teaching assisant for course “Probabilistic Artificial Intelligence”.
During my master I was a research intern at Computer Vision and Learning Grouph, hosted by Siyu Tang. I completed my semester project supervised by David Scaramuzza with distinction in the topic of event camera.
Before this, I received the Best Research Paper Award in CPHS with my graduation thesis and obtained Bachelor Degree in Automation at Shanghai Jiaotong University.
Recent Work
DE-NeRF: Deformable Neural Radiance Fields using RGB and Event Cameras(ICCV 2023)
Reconstruct fast moving/deformable objects using additional events supervision.
AMZ driverless perception core member
Event-YOLO, Data pipeline, model training and implementation, camera and lidar sequential fusion and KPI analysis.
Research assistant in Computer Vision and Learning Group
Human Pose Estimation from Egocentric Interaction Videos.
Semester project in Robotics and Perception Group
Calibrating an event-based structured light system.
Teaching assistant for Probabilistic Artificial Intelligence course
Segmentation using Bayesian neural network
Computer vision course project
Model fitting and multi-stereo vision.
VAMR full grade course project
Visual odometry project.
Deep Learning for Autonomous Driving
Lidar Motion Compensation.
Planning and Decision Making for Autonomous Robots course project
Rocket path planning and control with dynamic obstacle avoidance.
3D vision course project
On the Robustness of Local Feature Inversion Techniques.
Advanced topics in control course project
An exploratory study in formation optimization.
Get In Touch
If you have any interest in my project or want to communicate, please don't hesitate to contact me.