/PDSR

Submission to the PIRM2018 Super Resolution Contest

Primary LanguagePythonOtherNOASSERTION

Team Objective

Move along PSNR - Per curve

1. Undergrad's tips and tricks:

    1. Mean shift per patch
    1. RGB shuffle

2. DownMSE loss

3. Architecture adjustment

4. Laplace form

5. Handle faces directly

Overall Progress

08/21 Tuesday:

A. RGBshuffle
    1. Fix directory saving for model testing
    1. Implement rgb random shuffle, random in channels
    1. Test & confirm the shuffling
    1. save the shuffle img patches for sanity check
    • pre-shuffle & post-shuffle for both lr, hr
    1. Check the rgb training results, compare with the previous results
    1. understand the function __get_patch()
    1. Analysis and understand RGB results difference in paper
Mean shift per patch
B. DownMSE loss
    1. Save the LR images from SR, outside program for sanity check
    • Decide to use BiCubic Downsampling
    1. Integrate to program, bicubic down-sampling
    1. Implement downMSE
C. others:
    1. Display previous running results F_per vs. PSNR score
    1. Reply Cynthia
    1. Draft office desk email
Readings & theoretical analysis
    1. Read about PSNR theory
    1. Analysis GAN loss in different scalers

  • save and load the approximator

PDSR Prior work

Submission to the PIRM2018 Super Resolution Contest

The models are to big to fit on github, so they can be found here: https://drive.google.com/drive/folders/13eSV6d5cIFG67MMZ1by54uJcaExd6FHK?usp=sharing