ADMM-TGV deconvolution algorithm for STEM-EELS maps

Deconvolution algorithm for 3D STEM-EELS maps

Deconvolution algorithm for 3D STEM-EELS maps with total generalized variation (TGV), non-negativity prior and residual balancing for automated and unbiased parameter choice. The algorithm employs a noise model for scintillation-based CCD detectors, as they are often used in EELS, and considers the binning of the detector to create spectra as well as noise correlations occurring on such detectors.

A detailled derivation and description of the algorithm is provided in:

Christian Zietlow and Jörg K. N. Lindner

An unbiased ADMM-TGV algorithm for the deconvolution of STEM-EELS maps

Ultramicroscopy (2025), Article 114159

https://doi.org/10.1016/j.ultramic.2025.114159