Skip to content

Implementation of four different deep learning models for super-resolution.

License

Notifications You must be signed in to change notification settings

eteskeredzic/keras-superresolution

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 

Repository files navigation

keras-superresolution

Implementation of four different deep learning models for super-resolution.

The implemented models are:

The models are trained on traffic signs (the Traffic Sign Recogntion Database - http://www.nlpr.ia.ac.cn/pal/trafficdata/recognition.html)

You can either use the pretrained models for super-resolution (although they only work on the TSRD), or train them again from scratch on a different dataset (like DIV2k).

Helper functions include some stuff for inference time measurement, metrics (PSNR and SSIM), and generating LR images. A cubify helper function is also included, which divides larger pictures into smaller ones (for training on different datasets).