Yoshihiro Midoh et al 2005 Meas. Sci. Technol. 16 477 doi:10.1088/0957-0233/16/2/021
Yoshihiro Midoh, Katsuyoshi Miura, Koji Nakamae and Hiromu Fujioka
Show affiliationsFor the purpose of developing a recognition system for fine line patterns in the LSI, a statistical optimization method of the Canny edge detector for scanning electron microscope (SEM) images is proposed. The analysis of variance (ANOVA) is applied to the estimation of parameters in the SEM image: the grey level in the background, contrast, the line width and the noise level. These parameters are used to optimize statistically the Canny parameters, a scale parameter of the smoothing filter and a binarization threshold, by using noise characteristics of the SEM image. The proposed method is applied to model images and a real SEM image to show its validity.
02.60.Pn Numerical optimization
85.30.De Semiconductor-device characterization, design, and modeling
85.40.Qx Microcircuit quality, noise, performance, and failure analysis
Issue 2 (February 2005)
Received 29 June 2004, in final form 19 October 2004
Published 21 January 2005
Yoshihiro Midoh et al 2005 Meas. Sci. Technol. 16 477
Maciej Błaszak 2005 J. Phys. A: Math. Gen. 38 1667
I Terasaki et al 1991 Supercond. Sci. Technol. 4 S397
S J Lee and J N Reddy 2004 Smart Mater. Struct. 13 1176
Meelis Sildoja et al 2008 Metrologia 45 11
Ch Rothleitner et al 2009 Metrologia 46 283
Craig Warner et al. 2003 ApJ 596 72
O Novotný et al 2009 J. Phys.: Conf. Ser. 192 012021
Lea F Santos and Lorenza Viola 2008 New J. Phys. 10 083009
Marco Valerio Battisti 2009 J. Phys.: Conf. Ser. 189 012005