Researchers at the University of California, Los Angeles, have revealed a new advancement in optical imaging technology that could greatly improve systems for visual information processing and communication. The study was published in the journal Advanced Photonics Nexus.
The unidirectional diffractive processor transmits high-quality images in the forward propagation direction, represented with the blue line, from the field of view (FOV) A to FOV B, while effectively blocking the image formation in the backward propagation direction, represented with the brown line, from FOV B to FOV A. Image Credit: Ozcan Lab, UCLA
Based on partially coherent unidirectional imaging, the new method provides a small and effective way to send visual data in one direction while preventing transmission in the other.
This cutting-edge technology, developed by Professor Aydogan Ozcan and his multidisciplinary team, is intended to selectively transfer high-quality images from field-of-view A to field-of-view B. When viewed from the opposite way, B to A, the images are purposefully distorted.
With new possibilities for controlling the processing and transmission of visual optical information, this asymmetric image transmission may have wide-ranging effects on domains such as optical communications, augmented reality, and privacy protection.
Unidirectional Imaging Under Partially Coherent Light
The new system attempts to solve the problem of regulating light transmission to allow clear vision in one direction while blocking it in the other, a problem in optical engineering.
The practical applicability of previous unidirectional wave transmission solutions has been limited by their frequent reliance on intricate techniques like temporal modulation, nonlinear materials, or high-power beams under fully coherent lighting.
On the other hand, this UCLA invention purposefully introduces distortion and lowers power efficiency in the reverse direction (B to A) while utilizing partially coherent light to produce high image quality and power efficiency in the forward route (A to B).
We engineered a set of spatially optimized diffractive layers that interact with partially coherent light in a way that promotes this asymmetric transmission. This system can work efficiently with common illumination sources like LEDs, making it adaptable for a variety of practical applications.
Aydogan Ozcan, Professor, University of California, Los Angeles
Leveraging Deep Learning for Enhanced Optical Design
A significant component of this advancement is the physical design of the diffractive layers that comprise the unidirectional imaging system using deep learning.
The UCLA team optimized these layers for partially coherent light, ensuring a phase correlation length exceeding 1.5 times the light's wavelength.
Even in situations when the light source exhibits different coherence characteristics, the system's dependable unidirectional picture transmission is guaranteed by this meticulous optimization. Each image has a polarization-independent design, is small, and spans less than 75 wavelengths axially.
The design approach employed deep learning techniques to suppress image creation in the backward direction and retain high diffraction efficiency in the forward direction.
The researchers showed that their system is resilient to variations in the coherence qualities of the light and operates consistently across various image datasets and illumination circumstances.
“The ability of our system to generalize across different types of input images and light properties is one of its exciting features,” said Dr. Ozcan.
In the future, the researchers intend to investigate additional types of illumination sources and expand the unidirectional imager to other spectrum regions, such as the visible and infrared ranges.
These developments may push the limits of sensing and imaging, opening up new possibilities for inventions and applications. For instance, in terms of privacy protection, the technology might be used to shield private data from unauthorized views. Similarly, this technique might be used to manage how information is shown to various viewers in augmented and virtual reality systems.
This technology has the potential to impact multiple fields where controlling the flow of visual information is critical. Its compact design and compatibility with widely available light sources make it especially promising for integration into existing systems.
Aydogan Ozcan, Professor, University of California, Los Angeles
An interdisciplinary team from the California NanoSystems Institute (CNSI) and UCLA's Department of Electrical and Computer Engineering carried out this study.
Journal Reference:
Ma, G., et al. (2024) Unidirectional imaging with partially coherent light. Advanced Photonics Nexus. doi.org/10.1117/1.apn.3.6.066008.