Phat Nguyen (Peter)

I'm currently a visiting researcher at the MIT Distributed Robotics Lab. I completed my Bachelor's in Computer Science with a minor in Mathematics at the University of Massachusetts Amherst, where I was a research assistant at the Computer Vision Lab under Prof. Erik Learned-Miller.

My research interest lies at the intersection of computer vision, machine learning, and robotics. My goal is to bridge the gap between high-level cognition and low-level planning and control, with applications for autonomous vehicles.

Email  /  Resume  /  LinkedIn  /  Github

peter nguyen's photo
Research

My area of research interest is in autonomous vehicles, computer vision, machine learning, 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.

Teaching
Sept. 2023,

TA for CS326: Web Programming, UMass

May 2023,

TA for CS240: Reasoning Under Uncertainty, UMass

Jan. 2023,

TA for CS389: Machine Learning, UMass (Received distinguished TA award)

Activities and Leadership

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 was also part of the UMass Rocket Team between 2021-2022.

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.