ROHSA (Regularized Optimization for Hyper-Spectral Analysis) was developped by the Hyperstars collaboration at Paris-Saclay University (IAS/CEA) to study the statistical properties of interstellar gas through atomic and molecular lines.

This code is a Gaussian decomposition algorithm designed to decompose any kind of hyper-spectral observations into a sum of coherent Gaussian. It is written in Fortran 90 and can be run on a single CPU. A user-friendly ROHSApy python interface can be used to run the code easily.

ROHSA-GPU, a GPU implementation of ROHSA is currently developped by Jérémy Besson and Nicolas Gac.