Data necessary for liberation and/or exposure analysis is frequently obtained from two-dimensional (2D) polished section images of mineral particles of specified size. This data overestimates the extent of liberation and/or exposure, and in the past an empirical correction (locking) factor has been used to give a better estimate of the extent of dispersed phase liberation and/or exposure. With such 2D data, exposed grain surface area is estimated from perimeter analysis of sectioned particles. Although many researchers have reported 2D mineral liberation analysis in terms of area grade and/or perimeter based grain exposure, comparison of 2D perimeter based liberation with three-dimensional (3D) mineral liberation analysis has not been reported. In this current research, perimeter based estimates of exposed grain surface areas from 2D polished section analysis of feed and products from HydroFloat (HF) flotation of a copper porphyry ore are evaluated from corresponding 3D high resolution X-ray microtomography (HRXMT) data. Results show that for mean dispersed phase grades (sulfide mineral content) greater than 10%, perimeter based polished section analysis provides a good approximation of the exposed grain surface area determined by HRXMT. For mean grades less than 10%, stereological correction of the 2D data is required for satisfactory estimation of the exposed grain surface area. The most significant stereological correction with respect to exposed grain surface area is for the 0% grade class ("liberated" gangue minerals). For example, in the case of the 417 x 212 mu m size class, 2D perimeter analysis indicates that about 96% of the particles have 0% grain area exposure whereas 3D analysis shows that in reality about 77% of the particles have 0% grain area exposure.