Ensemble convolutional neural network for classifying holograms of deformable objects

On the use of deep learning for phase recovery

Phase recovery (PR) refers to calculating the phase of the light field from its intensity measurements. As exemplified from quantitative phase imaging and coherent diffraction imaging to adaptive optics, PR is essential for reconstructing the refractive index distribution or topography of an object and correcting the aberration of an imaging system. In recent years, deep learning (DL), often implemented through deep neural networks, has provided unprecedented support for computational imaging, leading to more efficient solutions for various PR problems. In this review, we first briefly introduce conventional methods for PR. Then, we review how DL provides support for PR from the following three stages, namely, pre-processing, in-processing, and post-processing. We also review how DL is used in phase image processing. Finally, we summarize the work in DL for PR and provide an outlook on how to better use DL to improve the reliability and efficiency of PR. Furthermore, we present a live-updating resource ( https://github.com/kqwang/phase-recovery ) for readers to learn more about PR.

Hologram classification of occluded and deformable objects with speckle noise contamination by deep learning

Advancements in optical, computing, and electronic technologies have enabled holograms of physical three-dimensional (3D) objects to be captured. The hologram can be displayed with a spatial light modulator to reconstruct a visible image. Although holography is an ideal solution for recording 3D images, a hologram comprises high-frequency fringe patterns that are almost impossible to recognize with traditional computer vision methods. Recently, it has been shown that holograms can be classified with deep learning based on convolution neural networks. However, the method can only achieve a high success classification rate if the image represented in the hologram is without speckle noise and occlusion. Minor occlusion of the image generally leads to a substantial drop in the success rate. This paper proposes a method known as ensemble deep-learning invariant occluded hologram classification to overcome this problem. The proposed new method attains over 95% accuracy in the classification of holograms of partially occluded handwritten numbers contaminated with speckle noise. To achieve the performance, a new augmentation scheme and a new enhanced ensemble structure are necessary. The new augmentation process includes occluded objects and simulates the worst-case scenario of speckle noise.