Low Light Image Enhancement

Processing pipeline for our project.
The convolutional neural network architecture used in our design.

Modern digital cameras have issues producing high quality images in extreme low light conditions, often resulting in high noise and color loss. Conventional techniques that employ denoising algorithms or burst imaging help with regard to somewhat low-light conditions but often fail to produce quality images in extreme low light conditions (<0.1 lux). We aimed to apply the concepts presented in this paper in order to demonstrate how CNNs can be applied to outperform these conventional algorithms.

Deliverables

Poster

Final Write-Up

Code

TA Check-in content

Initial Project Outline (Week 1)

Initial Reflection (Week 2)