Skip to content

Latest commit

 

History

History
executable file
·
5 lines (3 loc) · 683 Bytes

File metadata and controls

executable file
·
5 lines (3 loc) · 683 Bytes

Understanding-Experimental-Images-by-Identifying-Symmetries-with-Deep-Learning

This repository is designed for sharing source codes and tutorial notebooks for the symmetry identification project. The first part is the generation codes for the Imagenet Symmetry dataset, Atom Symmetry dataset, and Noise Symmetry dataset. The second part is benchmark notebooks and results using symmetry datasets. The Third part includes various methods to improve the ability of symmetry identification with deep learning models.

Symmetry datasets are available at https://zenodo.org/records/7384734. Due to the size limit, full-size datasets (10 million images) are available upon request.