Quantitative phase imaging in digital holographic microscopy based on image inpainting using a two-stage generative adversarial network

Abstract

Based on the hologram inpainting via a two-stage Generative Adversarial Network (GAN), we present a precise phase aberration compensation method in digital holographic microscopy (DHM). In the proposed methodology, the interference fringes of the sample area in the hologram are firstly removed by the background segmentation via edge detection and morphological image processing. The vacancy area is then inpainted with the fringes generated by a deep learning algorithm. The image inpainting finally results in a sample-free reference hologram containing the total aberration of the system. The phase aberrations could be deleted by subtracting the unwrapped phase of the sample-free hologram from our inpainting network results, in no need of any complex spectrum centering procedure, prior knowledge of the system, or manual intervention. With a full and proper training of the two-stage GAN, our approach can robustly realize a distinct phase mapping, which overcomes the drawbacks of multiple iterations, noise interference or limited field of view in the recent methods using self-extension, Zernike polynomials fitting (ZPF) or geometrical transformations. The validity of the proposed procedure is confirmed by measuring the surface of preprocessed silicon wafer with a Michelson interferometer digital holographic inspection platform. The results of our experiment indicate the viability and accuracy of the presented method. Additionally, this work can pave the way for the evaluation of new applications of GAN in DHM.

Quantitative Phase Imaging: Recent Advances and Expanding Potential in Biomedicine

“…In practice, these can result due to deviations in focal plane when screening across large surfaces or using multiwell plates, when changing media levels over the course of an experiment, or when using a previously acquired phase reference to enable faster imaging. Although other QPI methods like DHM can be improved with similar fitting procedures, background fluctuations (when present) and an inherent amount of both spatial and temporal noise due to the recovery of phase by numerical integration can impact cell segmentation. Overall, however, this approach can achieve high accuracy for measurements of the dry mass of cells, even at high cell densities …”

URL generative recognition method based on generative countermeasure network

“…The generative antagonistic network is composed of the generator network G and the discriminator network D . The optimization objectives of the two are completely opposite, and they are trained alternately to antagonize each other to achieve the optimum [7]. The normalized character vector is mapped to the input encoded data as the input vector of the GAN [8].…”

Multi-Class Classification and Multi-Output Regression of Three-Dimensional Objects Using Artificial Intelligence Applied to Digital Holographic Information

“…The CNN is one of the major deep neural networks and a basic building block of deep learning. The ability of a CNN to handle greater numbers of convolutional layers and pooling layers in the feature-extraction stage and process n number of neurons in the classification-layer stage has made its use feasible for different kinds of tasks such as classification, autofocusing, fringe pattern denoising, image segmentation, image super-resolution, and hologram reconstruction in digital holography [ 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 ]. A CNN [ 44 ] consists of feature-extraction and classification layers.…”