Large depth-of-field ultra-compact microscope by progressive optimization and deep learning.

Area of Science:

• Optics and Photonics
• Biomedical Engineering
• Computational Imaging

Background:

• Traditional optical microscopes are large, expensive, and offer limited performance.
• Miniaturization of high-performance optical systems is a significant challenge in imaging technology.

Purpose of the Study:

• To develop a highly integrated microscope with superior optical performance and a significantly reduced form factor.
• To demonstrate the feasibility of a compact microscope for portable diagnostic applications.

Main Methods:

• A progressive optimization pipeline was employed to systematically optimize aspherical lenses and diffractive optical elements.
• A simulation-supervision deep neural network was designed for spatially varying deconvolution during optical design.
• The integrated optics were combined with computational optics and deep learning techniques.

Main Results:

• Achieved optical performance exceeding a commercial microscope with a 5×, NA 0.1 objective in a 0.15 cm³ and 0.5 g package.
• Demonstrated over 10 times improvement in depth-of-field compared to traditional microscopes.
• Showcased successful integration into a cell phone for portable diagnostics with a wide variety of samples.

Conclusions:

• The proposed method offers a new framework for designing miniaturized, high-performance imaging systems.
• The integrated microscope enables advanced optical capabilities in a portable, accessible format.
• This technology has potential applications in remote and point-of-care diagnostics.