2025 Internship Project Overview

A recent advancement in deep convolutional networks, Deep Image Prior (DIP), uses the intrinsic structure of a neural network to enhance low-resolution images into higher-resolution images, a task known as super resolution. This project aims to generalize the DIP model to produce super-resolution geospatial maps of the East-Taylor watershed in Colorado. We used a low-resolution flood map of the area and several variables about the hydrology and terrain as a baseline for the network’s super resolution. The DIP-generated results showed good agreement when compared to our target high-resolution flood maps. This demonstrates that a deep learning approach can be used to generate a high-resolution geospatial map without requiring many training datasets, providing a more data-efficient alternative compared to traditional machine learning methods.

ORGANIZATION/RESEARCH GROUP

Integrated GroundWater Modeling Center (IGWMC), Department of Civil and Environmental Engineering, Princeton University

LOCATION

Princeton, New Jersey

 

Interns

Nathan Kang ’28
Major: Computer Science
Minor(s): Applied and Computational Mathematics, Statistics and Machine Learning
"I learned how to work with parallel computing architectures, GPUs and basic CUDA instructions, which were all new concepts and skills. In addition, I continued to hone my machine learning experience by adapting pre-existing code to a completely new application, which was both eye-opening and fun. We took an algorithm from image processing and generalized it for hydrological applications, which was very exciting and opens pathway to new projects. Overall, this internship further solidified my plans of going into academia and grad school, and it was a tremendous pleasure to work with my mentors."

Click to view the end-of-project presentation

 

Reed Maxwell
William and Edna Macaleer Professor of Engineering and Applied Science, Professor of Civil and Environmental Engineering and the High Meadows Environmental Institute

Mentors

  • George Artavanis, Research Software Engineer II, Research Computing
  • Xiao Xia Liang, Postdoctoral Research Associate, Civil and Environmental Engineering
  • Mia Peeples, Ph.D. Candidate, Civil and Environmental Engineering
  • Marie Joe Sawma, Ph.D. Candidate, Civil and Environmental Engineering