“All-in-One” Method Reconstructs Holograms Using Deep Neural Network

Digital holography is a widely used imaging technique that can record the entire wavefront information, including amplitude and phase, of a 3D object. With an interferometer and an image sensor, a 2D hologram can be acquired and stored in a computer. Due to its noninvasive and label-free properties, holography has been applied to biological imaging, air/water quality monitoring, and quantitative surface characterization measurement.

After capturing a digital hologram, appropriate algorithms are used to reconstruct the object numerically. Conventional approaches require prior knowledge and cumbersome operations for an in-focus and successful reconstruction. For a 3D object, an all-in-focus image and a depth map are particularly desired for many applications, but conventional reconstruction methods tend to be computationally demanding.

Recently, researchers at the University of Hong Kong demonstrated an automated "all-in-one" method that can tackle holographic reconstruction problems through a deep neural network trained with appropriate data. After appropriate training, the network can holographically reconstruct the amplitude, quantitative phase, extended focused image, and depth map. The cumbersome operations involved in conventional reconstruction approaches are avoided and system parameters become unnecessary. The intensive computational demand is also significantly alleviated by total automation. Qualitative visualization and quantitative measurements confirm the superior performance of the learning-based method over conventional ones.

Through this data-driven approach, it is possible to reconstruct a noise-free image that does not require any prior knowledge and can handle diverse reconstruction modalities simultaneously. To advance this work, the researchers plan to apply this technique to high-speed and high-resolution temporal holographic reconstruction of 3D scenarios. They note that this method is universal to various digital holographic configurations and is potentially applicable to biological and industrial applications.

(a) Schematic of the deep learning workflow and the structure of HRNet. It consists of three functional blocks: input, feature extraction, and reconstruction. In the first block, the input is a hologram of either an amplitude object (top), a phase object (middle) or a two-sectional object (bottom). The third block is the reconstructed output image according to the specific input. The second block shows the structure of HRNet. (b) and (c) Detailed structures of the residual unit and the sub-pixel convolutional layer, respectively.

Read the original research article in the open-access journal Advanced Photonics. Z. Ren, Z. Xu, and E.Y.M. Lam, "End-to-end deep learning framework for digital holographic reconstruction," Adv. Photonics, 1(1) 016004 (2019).