Current and future Cosmic Microwave Background (CMB) experiments target the measurement of the B-mode polarised signal, induced by primordial inflationary gravitational waves. Polarised Galactic foregrounds can bias this measurement and cause a false detection. Here we describe a robustness test for the validation of an eventual detection, using scattering transforms as statistical descriptor of component separated CMB maps. The proposed test is able to identify a false detection in ~90% of our simulations, in a realistic configuration.
