Algorithms for metalens & hologram, incl. Deep learning

Algorithms for metalens & hologram, incl. Deep learning [DL]


● Research fields

       Imaging optics,

       Diffraction optics, 

       Micro-nano optics, 

       Deep learning.


● Research details

Metalens, an ultrathin and planar nanostructure with flexible control of the amplitude, phase, and polarization of light, has been demonstrated to replace bulky optical elements in many applications, such as imaging, spectroscopy and holography.

Typically, trial and error approaches are adopted in the metalens design where a progression of commercial software simulations that iteratively solve Maxwell’s equations. However, such full-wave simulation method is really time-consuming. On the other hand, a standard procedure of metalens design is also required for speeding up this process. This project includes the following two sub-researchs.

Sub-research 1: Deep learning enabled metalens design [DL1]

Deep learning is a class of machine learning techniques that use multilayered artificial neural networks for automated analysis of signals or data. Recently, deep neural networks, such as deep convolution neural network (CNN), have been successfully applied to solve numerous imaging-related problems including optical microscopy, computed tomography and so on.

This research is targeting for applying deep learning to design metalens for spectral and imaging device. We envision that a combination of such model with deep learning design method can expedite the metalens design with the expected performances.

A SiN mealens array for full-color integral imaging

       [1] Metalenses at visible wavelengths: Diffraction-limited focusing and subwavelength resolution imaging, Science 352, 1190-1194 (2016).
       [2] A broadband achromatic metalens array for integral imaging in the visible, Light Sci. Appl. 8, 67 (2019).
       [3] Deep learning for accelerated all-dielectric metasurface design, Optics Express 27, 27523-27535 (2019).

Sub-research 2: Software Development for Metalens Design and Hologram [DL2]

Computer-generated holography (CGH) is the method of digitally generating holographic interference patterns. A holographic image can be generated by illuminating the holographic interference pattern using a suitable coherent light source. Metalens is composed of subwavelength-spaced phase nanopillars at an interface, which can control the properties of light, such as amplitude, phase and polarization. Metalens focuses the light into a focal plane with subwavelength focal spot for high-resolution imaging. RCWA, FDTD and field tracing methods are frequently used to metalens design.

This task will develop your skills to write your own code in hologram and metalens design and support the software development and testing at the Metaphotonics. This research is to achieve efficient information processing and you are encouraged to bring your own ideas.

  

3D holographic image (left) and a metalens (right)

       [1] Viewing-angle enlar gement in holographic augumented reality using time division and spatial tiling," Optics Express 21, 12608 (2013).
       [2] Dynamic holographic imaging of real-life scene, Optics and Laser Technology 119 (2019) 105590.
       [3] Silicon Nitride Metalenses for Close-to-One Numerical Aperture and Wide-Angle Visible Imaging, Phys. Rev. Appl. 10, 014005 (2018).                                                                                                                           


● Required background

      Electrodynamics, Fourier optics, Applied optics, Information optics, Computational Physics, Programming skill (Matlab, Python and so on).


● Research suitability

       PhD, MPhil, Undergraduate [Position DL1 and DL2].


● Contact supervisor

        DL1:  Associate Professor Rui Chen (硕导) chenr229.at.SYSU

        DL2:  Professor Jian-Wen Dong (博导) dongjwen.at.SYSU