Intensity Mapping of line fluctuations is a relatively new probe in astronomy, where tomographic large-scale maps provide us with information on a broad range of scales and redshifts. Due the potential of exploiting and combining multiple lines as tracers, intensity mapping enables us to map structures from high redshifts of the Epoch of Reionisation to present times, thus informing about structure growth as well as astrophysical processes. This multi-line and large-scale imaging of intensity fluctuations is an ideal candidate for the application and development of deep learning and computer vision techniques, that have the potential for optimal treatment of such imaging. I will review recent developments in both fields, highlight how line intensity maps are modelled, as well as present source detection and parameter inference in preparation for 21cm measurements with the SKA (Square Kilometre Array). Finally, I will sketch our road ahead with machine learning-based analysis layers for surveys, briefly showcasing deblending of galaxy imaging, as well as object classification.