Mohammed Ouali
Noise model, Noise contamination, SNR, PSNR, Autocorrelation
Very few noise removal methods have been developed for scanning electron microscopy (SEM), and most of the noise removal algorithms come from standard image processing. These algorithms are designed for a certain type of image formation process, very often optical. Moreover, these conventional algorithms are not always easy to use by SEM operators because of their number of parameters (size, shape, weight, and the number of iterations to name a few). Furthermore, the setting of these parameters requires a sound understanding of the underlying algorithm. It is then very important to devise an adequate noise removal filter for such an image formation process (SEM). However, it is still unclear how noise affects SEM images. In this contribution, we tackle the characterization of noise in SEM imagery. A noise identification taxonomy is suggested as well as a method for off-line SNR estimation.
Important Links:
Go Back
