ganX -- generate artificially new XRF a python library to generate MA-XRF raw data out of RGB images

In the last decade, we have witnessed a truly remarkable rise of Artificial Intelligence, Statistical and Deep learning methods (for a non exhaustive list of papers on the history of deep learning, see [1, 2, 3], and references therein). Inspired by the incredible results obtained thanks to the application of such methods to scientific problems, its adoption in the field of nuclear imaging applied to Cultural Heritage (CH) has begun (see, e.g., [4, 5, 6, 7, 8], especially the nice overview [9], and, of course, the references therein), also in the field of X-ray fluorescence Macro mapping (MA-XRF) [10, 11, 12, 13, 14, 15]. In MA-XRF, the imaging apparatus produces a data cube which, for each pixel, is formed by a spectrum containing fluorescence lines associated with the element composition of the pigment present in the pictorial layers. MA-XRF data cubes offer an ideal framework for application of unsupervised statistical learning methods [13], due to the huge number of pixel XRF histogram w.r.t. the relatively small number of employed pigment palettes. Unfortunately, in the realm of supervised statistical (deep) learning applied to CH-based analysis, the situation is flipped [14], since the data cube production is slow, obtaining a small dataset for the complexity of the various task at hand (like automatic pigment identification [14, 15], element recognition [16], and even colour association [11, 12]). This justifies the emphasis put on the creation of ad hoc synthetic MA-XRF dataset [14].