Aman Agarwal

About Me

Sc.M. @ Brown University
Visual Computing & AI

amanag@brown.edu
Github
LinkedIn
💻 Watch me in 3D


I’m a Master’s student in Computer Science at Brown University, specializing in Visual Computing and AI. I am advised by Dr. James Tompkin, and my current research focuses on dynamic scene reconstruction for rendering.

My primary interests lie in Computer Vision and Graphics, with a focus on Inverse Rendering via Neural Fields, including NeRFs and Gaussian Splatting.

Experience

2025

Graduate Researcher : Dynamic Gaussian Splatting & NeRFs

Novel View Synthesis for Dynamic Scenes

2024

Research Intern : Augmented Gaussian Splatting & NeRFs

Conducted ablation studies FSGS - Real-Time Few-Shot View Synthesis using Gaussian Splatting to calculate the efficiency of priors used in the paper and researched on effects of augemented networks in NeRFs & Gaussian Splatting

2023

Research Intern : Integrated NeRF-Capture into Zip-NeRF

Integrated NeRFCapture, an app that use Apple's ARKit to computer poses of photos simultaneously as they captured, into the (unofficial) Zip-NeRF pipeline to study accuracy of pre-computed poses over COLMAP and attempting to reduce entire pre-processing time to over 50%.

Projects

Volummetric Rendering for Clouds & Procuderal Terrain, on OpenMP

Designed and implemented a custom ray-marcher and ray-tracer from scratch to render realistic clouds and generate procedural terrain. The rendered scene includes dynamic interactions with point lights, creating visually accurate and immersive effects.

Neural Radiance Fields

Implemented NeRFs

Implemented the Neural Radiance Fields paper by Ben Mildenhall et al.; Removed heirarical sampling to reduce complexity while maintaining PSNR scores.

Neural Fields Thumbnail

Neural Fields with Encoding

A comparison of different encoding methods for 2D inputs

Loving Vincent Thumbnail

Loving Vincent

Convert your videos into moving paintings of your choice

Rescue Shark Thumbnail

Rescue Shark

Built an SOS application for Truck Drivers - Won a MLH Hackathon for "Most Creative Use of Twilio"


Template Credits : Matias Turkulainen