Learning-based method to reconstruct complex targets through scattering medium beyond the memory effect.

Enlai Guo,Jing Han,Chao Zuo,Shuo Zhu,Lianfa Bai,Yan Sun

doi:10.1364/oe.383911

Abstract

Strong scattering medium brings great difficulties to image objects. Optical memory effect makes it possible to image through strong random scattering medium in a limited angle field-of-view (FOV). The limitation of FOV results in a limited optical memory effect range, which prevents the optical memory effect to be applied to real imaging applications. In this paper, a kind of practical convolutional neural network called PDSNet (Pragmatic De-scatter ConvNet) is constructed to image objects hidden behind different scattering media. The proposed method can expand at least 40 times of the optical memory effect range with a average PSNR above 24dB, and enable to image complex objects in real time, even for objects with untrained scales. The provided experiments can verify its accurateness and efficiency.

Highlights

Scattering medium exists generally in biological tissues and is the main interference source in the field of astronomical imaging
The typical ones include optical coherence tomography, wavefront modulation, image reconstruction based on transmission matrix, and reconstruction based on point spread function (PSF)
The FOV of optical memory effect (OME) has been successfully expanded through several novel techniques, but most of them belong to invasive method, which rely on priori knowledge[6-8].As far as we know, the best non-invasive method to imaging beyond the OME range, is a double loop iterative algorithm proposed by Dai[9], which can restore two targets respectively whose total size is beyond OME, the algorithm can expand at least three times of OME scope shown by experiments, no prior information is required, but there are still some constraints for the target to be restored : the target must be in regular shape and in two independent OME region

Summary

Introduction

Scattering medium exists generally in biological tissues and is the main interference source in the field of astronomical imaging. Tian et al constructed a CNN network 11 with the "one-to-all" reconstruction capability of single optical statistical characteristics, which can reconstruct the speckle image generated by untrained ground glass, only condition is that untrained scattering medium has the same statistical characteristics as the ground glass used in making the training set. Breaking through the limitation of OME on the FOV, good reconstruction ability for complex targets and strong generalization ability, are the basic conditions for imaging method through scattering media to be practical. A kind of convolutional neural network named PDSNet (Pragmatic De-scatter ConvNet) is constructed to break through the limitation of OME on FOV in a data-driven way, combining with the principle of traditional speckle correlation imaging algorithm to guide the design and optimization of the network. PDSNet is a neural network structure suitable for random scales and complex targets

Basic theory

Simulation

Objective

Real System Experiment

Conclusion