Inverted EEMD: a robust method to identify narrow absorption features form spectral data and cubes

Extracting information from complex data is a challenge shared by multiple frontiers of modern astrophysical research. Among those, analyzing spectra cubes, where the emission is mapped in the position-position-velocity space is a difficult task given the vast amount of information contained within. The cubes often contain a superposition of emissions and absorptions, where extracting absorption signatures is often necessary. One example is the extraction of narrow absorption structures in HI 21 cm emission spectra. These HI self-absorption (HISA) clouds trace the cold HI gas in interstellar space. We introduce an automatic and robust method called the \emph{inverted EEMD} method to extract narrow features from spectral cubes. Our method is based on the EEMD method, an established method to decompose 1d signals. The method is robust and parameter-free, making it useful in analyzing spectral cubes containing localized absorption signals of different types. The inverted-EEMD method is suitable for the analysis of spectral cubes where it can produce a cube containing the absorption signal and one containing the unabsorbed signal, where cold clouds can be identified as coherent regions in the absorption map. A Python implementation of the method is available at \url{https://github.com/zhenzhen-research/inverted_eemd_map}.