prismPYP: Power-spectrum and image domain learning for self-supervised micrograph evaluation

prismPYP implements a SimSiam-based self-supervised pipeline for classifying cryo-EM micrographs using both real-space and Fourier-space features.

The goal is to automatically uncover image-quality categories such as vitreous ice, crystalline ice, contaminants, and support film entirely without labels.

The framework builds on a PyTorch implementation of SimSiam, distributed under the Attribution-NonCommercial 4.0 International license.

The software is developed and maintained by the Bartesaghi Lab at Duke University and released under the BSD 3-Clause license.