DISCO: Deprojection Image Software for Circumstellar Objects

An open-source software package for the automated and interactive analysis of protoplanetary disks from astronomical FITS observations, primarily focused on ALMA continuum data.

The project combines machine learning, numerical optimization, and visualization tools to estimate disk geometry parameters such as inclination, position angle, center offsets, inner/outer radii, and radial intensity profiles. It also supports multi-band analysis, beam homogenization, residual inspection, and FITS/CSV export workflows.

DISCO includes DiscoNet, a CNN trained on both synthetic datasets and realistic CASA interferometric simulations to improve performance on real observations.

Built with Python, FastAPI, React, PyTorch, and the Astropy scientific ecosystem.

Developed within the YEMS Millennium Nucleus under the supervision of Sebastián Pérez and Camilo González-Ruilova.

Links