Phat Nguyen (Peter)

I am Peter, a senior at the University of Massachusetts Amherst studying Computer Science and Mathematics.

Currently, I am a research assistant at the UMass Computer Vision Lab under Professor Erik Learned-Miller. I am currently working on estimating optical flow that are compatible with camera translation. Starting this summer, I am involved in autonomous vehicles research at the MIT Distributed Robotics Lab (DRL). My area of work is in simulation engines and large language models to train full-scale autonomous vehicles.

Previously, I co-founded Stella Agritech to create technologies that enhance the efficiency of agriculture. I also led ISHCMC's sustainability program with Tanya Meftah, which was nominated by the European Chambers of Commerce for the Best Sustainable Business Initiative in Vietnam in 2019. I hope to research ideas that will evolve out of the lab and turn into great companies.

Email  /  Resume  /  LinkedIn  /  Github

peter nguyen's photo
Research

My area of research interest is in autonomous vehicles, computer vision, machine learning, multi-agent, and robotics.

Robust Frame-to-Frame Camera Rotation Estimation in Crowded Scenes
Fabien Delattre, David Dirnfeld, Phat Nguyen, Stephen Scarano, Pedro Miraldo, Michael Jeffrey Jones, Erik Learned-Miller
International Conference on Computer Vision (ICCV), 2023
project page / poster / arXiv

We introduce a novel generalization of the Hough transform on SO3 to efficiently find the camera rotation most compatible with the optical flow. We also provide a new dataset of 17 highly dynamic video sequences called BUsy Street Scenes (BUSS).

Classifying Facial Expressions using Vision Transformers and Convolution Neural Networks
Peter Nguyen, Peter Phan
poster

Comparing predictive performances between ViT and CNNs to classify facial expressions.

Activities, and Extracurriculars
UMass Aerospace and Rocket Team, NASA SLI 2022, Payload Software Engineer

Building the onboard flight inertial reference system (IRS) to continuously compute the launch vehicle's dead reckoning position in a GPS-denied environment with a 75 meter (250 ft) resolution


   


Some inspiration from Jon Barron.