Ventura

Learning Panoramic View Synthesis

A synthetic data generation and evaluation pipeline for panoramic computer vision models. This system extracts image and depth data from 3D datasets and outputs rendered comparisons. The evaluation pipeline gives researchers the ability to generate highly accurate, novel snapshots from multiple 3D environment datasets, lowering the resource cost of experimentation.

Learning to Denoise Low-Dose CT Scans

This SURP project explores the use of deep learning to denoise raw CT image data acquired at low-radiation dose. The goal is to recover clean CT images from noisy data that are comparable to CT scans imaged at full-radiation dose.

Coronavirus Update and Resources