In performing SEM Critical Dimension (CD) measurements on photomasks in dense line and space arrays it is often difficult to distinguish between whether a feature is a line or space. This is a result of tone shifts that occur affecting contrast on target images. The inability to reliably differentiate lines and spaces leads to the inclusion of fliers, or inaccurate measurements into automated measurement results. In an effort to overcome this phenomenon a new algorithm has been developed to increase the robustness of the CD SEM measurements to insure reliable data acquisition. This new algorithm takes into account apparent tone reversals on a variety of today's photomask material types. This paper will detail the various elements of the new algorithm and show before and after test results of improved recognition performance.