Signal-to-noise ratio estimation for SEM single image using cubic spline interpolation with linear least square regression

K. S. Sim, F. F. Ting, J. W. Leong, C. P. Tso

Research output: Contribution to journalArticleResearchpeer-review

10 Citations (Scopus)


A novel technique based on cubic spline interpolation with linear least square regression (CSILLSR) is developed to calculate the signal-to-noise ratio (SNR) of scanning electron microscope (SEM) images. The SNR from CSILLSR method is compared with methods of nearest, linear interpolation, a combination of linear interpolation and nearest, nonlinear least square regression, autocorrelation Levinson-Durbin recursion, and adaptive slope nearest neighbourhood. Samples of SEM images with various accelerating voltages, beam diameters, surface tilts and contrast were applied to evaluate the performance of CSILLSR method in terms of SNR values of the SEM images. The new method is able to generate more accuracy results than the other six methods. In addition, the CSILLSR Wiener filter appears to be the best filter to reduce and remove white Gaussian noise from SEM images as compared to the average filter and median filter.

Original languageEnglish
Pages (from-to)151-165
Number of pages15
JournalEngineering Letters
Issue number1
Publication statusPublished - 1 Feb 2019
Externally publishedYes


  • Cubic spline interpolation
  • Gaussian noise
  • Linear least square regression
  • SNR estimation

Cite this