The 'weak scatterer' approximation is usually inappropriate for re-entrant target structures, and ISAR images based on this approximation often display unwanted 'artifacts' which can complicate image interpretation. Last year, we presented an image restoration method which can be used to mitigate these artifacts without significant impact on neighboring image components and demonstrated its application using anechoic chamber data (SPIE Proc., Radar Processing, Technology, and Applications II, vol. 3161, pp. 9-19 (1997)). Since then, we have further examined this novel filtering technique and applied it to measured ISAR data. In addition, duct shape-specific parameters can be recovered using this analysis, and we have looked at the possibility of applying these to the problem of radar-based target classification. Our presentation will review the past theory and present results of our investigations over the past year.