Wavelet Based Thresholding for Fourier Ptychography Microscopy

Computational microscopy algorithms can be used to improve resolution by synthesizing a bigger numerical aperture. Fourier Ptychographic (FP) microscopy utilizes multiple exposures, each illuminated with a unique incidence angle coherent source. The recorded images are often corrupted with background noises and preprocessing improves the quality of the FP recovered image. The preprocessing involves data denoising, thresholding and intensity balancing. We propose a wavelet-based thresholding scheme for noise removal. Any image can be decomposed into its coarse approximation, horizontal details, vertical details, and diagonal details using suitable wavelets. The details are extracted to find a suitable threshold, which is used to perform thresholding. In the proposed algorithm, two wavelet families, Daubechies and Biorthogonal with compact support of db4, db30, bior2.2 and bior6.8, have been used in conjunction with ptychographic phase retrieval. The obtained results show that the wavelet-based thresholding significantly improves the quality of the reconstructed FP microscopy image.