Efficient digital holographic 3d human image reconstruction and improvement on mobile devices

Abstract

Advanced digital holography attracts a lot of attentions for 3D visualization nowadays. The representation of high-resolution digital holographic 3D human images suffers from computational inefficiency on the mobile devices due to the limited hardware for digital holographic processing. Specifically, to reconstruct the high-quality holographic image needs to compensate for the phase aberration, which needs lots of expensive optical hardware components to acquire measurements such as different axial distances, illumination angles, wavelengths, polarization states, and so on. To reduce computational complexity in digital holographic 3D human image reconstruction, we propose an efficient and effective algorithm to simplify Fresnel transforms for the mobile devices. Our algorithm reduces the number of FFTs and fastens the calculation of the exponential function in the Fresnel integral for the digital holography image reconstruction. Specifically, we reformulate the Fresnel integral and use a polynomial approximation to approximate the exponential function. In the holographic image quality improvement, we modify a maximum a posteriori (MAP) estimation to improve the quality of the reconstructed holographic 3D image restoration. Our algorithm outperforms previous approaches in not only smaller running time but also the better quality of the digital holographic 3D human image representation for the mobile devices.