Adaptive spatial filtering based on fuzzy C-means and phase in off-axis digital holographic microscopy

Digital holographic microscopy can quantitatively image the biological samples label-free and noninvasively. It is key to extract the +1-term spectrum from the hologram spectrum, which is crucial to the quality of the reconstructed image. Therefore, an adaptive spatial filtering method based on fuzzy C-means and phase spectrum of a hologram is proposed to extract the +1-term spectrum without any prior knowledge. The maximum phase value point of phase spectrum is found, which must be located in the +1-term spectrum. Then, this point is first introduced to locate the +1-term spectrum region. Two classifications and three regions (+1-term, −1-term, and zero-order term spectra regions) are obtained by fuzzy C-means in the amplitude spectrum. Subsequently, the minimum distance between the centroids of the three regions and the maximum phase point is used to judge the +1-term spectrum region. Finally, a filtering window is obtained by the edge of the +1-term spectrum region and the +1-term spectrum is adaptively extracted. Compared to other spatial filtering methods, the proposed method avoids dependence on a prior custom mask and suppresses the higher frequency noise. Most importantly, the experimental results on a number of human cells and a phase step demonstrate the feasibility and efficiency of the proposed method.