In most proximity effect correction schemes, circuit features are assumed to be rectangular. However, there are many circuit patterns which contain nonrectangular shapes of features such as circles, circular arcs, polygons, etc. In this article, efficient ways to handle such features for exposure estimation required in proximity effect correction are proposed. A hierarchical approach, where a nonrectangular feature is partitioned into correction shapes first, and then each correction shape may be further decomposed into exposure shapes if necessary, is taken in order to develop efficient and generally applicable schemes. Specifically, for exposure estimation. four schemes (direct, slicing, hybrid, and coordinate transformation methods) are described. of which performances have been analyzed through an extensive simulation. (C) 2004, American Vacuum Society.