Skip to content

Low light image/video enhancement based on reference-less loss functions

Notifications You must be signed in to change notification settings

kennethmitra/DIP-LowLightEnhancement

Repository files navigation

DIP-LowLightEnhancement

Link to our presentation describing the project: https://www.youtube.com/watch?v=owQkejcJhro

File Descriptions

Model.py - Model outputting pixelwise RGB curves and enhanced image

LossFunctions.py - Defines functions based on various metrics to compute loss of an enhanced image without a reference image

Train.py - Script used to train the model

Test.py - Script used to evaluate model on test data

CreateFrames.py - Splits a video into frames

EnhanceFrames.py - Applies low light enhancement to frames of video (using several versions of the model + an image quality net to select the best image)

CreateVideo.py - String together enhanced frames into a video

About

Low light image/video enhancement based on reference-less loss functions

Topics

Resources

Stars

Watchers

Forks