Skip to content

Files

Latest commit

708a116 · Apr 20, 2023

History

History
26 lines (22 loc) · 1.55 KB

README.md

File metadata and controls

26 lines (22 loc) · 1.55 KB

Deep-Learning

Repository for the Deep Learning reproducability project. It is based on the Learning to See in the Dark paper.

Scripts

This repository contains several scripts.

  • Testing: use this to test your own images on a trained model
    • Ensure you set the directories for both input, output and model correctly
    • Do not forget to change the file extention correctly as this is different depending on the camera
    • Do not forget to change the bit level of the camera
    • The code will automatically select the right black level and bayer filter arrangement
  • Training: use this for training your own model
    • Ensure you set the directories for input and output correctly
  • PSNR and SSIM: use this to calculate PSNR and SSIM values for pairs of output/ground truth images
    • Again, ensure you set the directories correctly
  • Plot loss: use this to plot the loss of your own trained models

Data

This repository also contains data of several trained models.