Noise Filtering Method of Digital Holographic Microscopy for Obtaining an Accurate Three-Dimensional Profile of Object Using a Windowed Sideband Array (WiSA)

Abstract

In the image processing method of digital holographic microscopy (DHM), we can obtain a phase information of an object by windowing a sideband in Fourier domain and taking inverse Fourier transform. In this method, it is necessary to window a wide sideband to obtain detailed information on the object. However, since the information of the DC spectrum is widely distributed over the entire range from the center of Fourier domain, the window sideband includes not only phase information but also DC information. For this reason, research on acquiring only the phase information of an object without noise in digital holography is a challenging issue for many researchers. Therefore, in this paper, we propose the use of a windowed sideband array (WiSA) as an image processing method to obtain an accurate three-dimensional (3D) profile of an object without noise in DHM. The proposed method does not affect the neighbor pixels of the filtered pixel but removes noise while maintaining the detail of the object. Thus, a more accurate 3D profile can be obtained compared with the conventional filter. In this paper, we create an ideal comparison target i.e., microspheres for comparison, and verify the effect of the filter through additional experiments using red blood cells.