CARMEN has been developed in MatLab and uses specialize toolboxes as Image Processing, Parallel Computing, and Image Acquisition. MatLab, though commercial, is widely used by the research community, which allows us to incorporate third party algorithms easily and facilitates multicenter collaboration to customize the tool.

CARMEN, as most medical applications, manages big data volumes. To avoid the possible slow-down caused by a too demanding use of RAM memory, CARMEN benefits from Vectorization and Parallelization paradigms, using GPU accelerated MatLab built-in functions to improve its performance. Besides, data representation is being carefully intended to be agile.

Ideally, the new GUI should run on a multi-CPU workstation with a dedicated graphic card, being MC simulations distributed in a remote cluster; nonetheless CARMEN can run in a simple desktop PC. CARMEN works on Linux and Windows.

The platform encompasses four working modes: load and edit, verification, planning and evaluation. The tool reads the DICOM objects associated with RT, as well as the most generalized formats in MC simulation and dose storage; besides, it provides tools to calculate the SUV, clean the CT, filters the data, edit the calibrator curve, and a powerful UI to edit and create structures, among others.