标题：Cubic Spline Hermite Interpolation with Linear Least Square Regression for Single Scanning Electron Microscope Image Signal-to-Noise Ratio Estimation
期刊名称：IAENG International Journal of Computer Science
印刷版ISSN：1819-656X
电子版ISSN：1819-9224
出版年度：2020
卷号：47
期号：4
语种：English
出版社：IAENG - International Association of Engineers
摘要：The quality of the scanning electron microscope (SEM) images can be estimated using the signal-to-noise ratio (SNR). SNR is defined as the ratio of the desired signal to background noise. The noise which appeared in the SEM image is called Gaussian noise. Thus, if the SNR value is high, the image will have better quality since there is more useful information (the signal) than unwanted data (the noise). However, existing SNR estimation methods such as Nearest Neighbourhood (NN), Linear Interpolation (LI), and combination of Nearest Neighbourhood and Linear Interpolation unable to provide satisfactory results in estimating the SNR value. So, to prevent the loss of important information of SEM images, the novel SNR estimation method named Cubic Spline Hermite Interpolation with Linear Least Square Regression (CSHILLSR) has been proposed and formulated. The proposed method is compared with existing methods in terms of absolute error of SNR values, F-test, and Student’s t-test. The result shows that the proposed method having a lower absolute error as compared to other methods and there is no significant difference between the actual and estimated SNR value at a 95% confidence level. This indicates that the proposed CSHILLSR able to provide better accuracy in estimation of SNR value as compare to the existing methods.
关键词：Gaussian noise;SEM image;Absolute error;Signal-to-noise ratio (SNR);F-test;Student’s t-test
