Simulated, uncorrupted images with 5 mas pixel size (left column), observed dirty images generated with CASA (middle) and the reconstructed low-resolution shape model (right). Note that the middle-column images are not needed in inversion; we use the direct FT data instead. The images are what would be seen if the raw data were deconvolved for viewing purposes as is usually done for ALMA targets. The test shape model is from Ostro et al. (2000).