Efficient Digital Hologram Generation using Position Clustering and FFT-based Structural Modification

In a digital hologram, 3D information of an object is recorded on a holographic fringe pattern obtained from the interference pattern between the object and reference waves, or the numerical diffraction equations. Numerical equation-based hologram generation is highly computationally complex because the relationship between a point of the object and all the positions on a fringe pattern need to be considered. In this work, we present an efficient method for digital hologram generation using position clustering and FFT-based structural modification by Taylor series approximation. To evaluate the performance of the proposed method, it was compared to the Rayleigh-Sommerfeld (RS) diffraction for two point clouds and the gray and depth images were used to generate a digital hologram. Experimental results show that the proposed method has much lower computational complexity in terms of the numerical structure while similar visual quality was preserved; the proposed method is adjudged to more suitable for parallel processing. Therefore, we believe that the proposed method can be a useful tool for numerical digital hologram generation.