Bilateral Filtering and Wavelets based Image Denoising: Application to Electron Microscopy Images with Low Electron Dose

Image denoising is a very important step in Cryo-transmission electron microscopy (cryo-TEM) and energy filtering TEM images, before the three-dimensional reconstruction (tomography reconstruction), because they normally have a problem of high noise level, which causes a loss in the contained information and difficult the process of alignment required for tomographic 3D reconstruction. This paper outlines the forcefulness of wavelets transform by using different wavelets family, Daubechies (which presented in our experiments by Haar, DB2 and DB3), Symlet (‘sym8’) at hard and soft thresholding with different levels of decomposition, plus the effectiveness of bilateral filter in one of the important tasks in image processing which is the de-noising. The two denoising methods are tested on real datasets of TEM images took at different time exposure. To assess our results, we’ve chosen the signal-to-noise-ratio SNR criterion beside the visual quality of the obtained images. Compared to Daubechies family we found that ‘sym8’ was the best in noise removal for these particular TEM images. Also, it has been shown through the obtained denoised images, that the bilateral filter is very effective by its capability to avoid blurring the high resolution details plus suppressing noise, therefore enhancing the visual quality of the TEM images. It seems better to answer the tradeoff between the use of Low electron doses to reduce the radiation damage, and feasibility to improve SNR after acquisition.