Spectrally Encoded Nonscanning Imaging through a Fiber

Ningzhi Xie, Quentin A. A. Tanguy, Johannes E. Fröch, K. F. Böhringer, Arka Majumdar

TL;DR

This paper introduces a novel method for imaging through a single optical fiber without mechanical scanning, utilizing spatial-spectral encoding with a microfabricated metasurface filter array. This technique allows for robust image acquisition even when the fiber is bent or moved, paving the way for more compact and flexible endoscopic imaging devices.

What Problem Does This Paper Address?

The increasing need for miniaturized and flexible imaging devices, particularly in fields like neuroimaging and microsurgery, has highlighted the potential of optical fibers. However, single-mode optical fibers can only transmit intensity and phase information for a single pixel, and transmitting an entire image typically requires bulky scanning mechanisms or complex fiber bundles with limited resolution. Existing methods for imaging through multi-mode fibers often lack robustness, requiring recalibration if the fiber is disturbed. This paper addresses the challenge of achieving high-resolution, non-scanning imaging through a single optical fiber by developing a robust spatial-spectral encoding technique.

Methods

The core of this work is a Spatial-Spectral Encoding (SSE) device fabricated using a metasurface integrated into a Fabry-Pérot (FP) cavity. Here's a breakdown of the key methods:

  • Metasurface Design: A 4x4 array of pixels was designed, where each pixel acts as a spectral filter. These filters are created by embedding Si₃N₄ square meta-atoms within an FP cavity. The lateral size of these meta-atoms is engineered to control the phase shift, which in turn dictates the resonant wavelength of the FP cavity. This allows for the creation of distinct, highly orthogonal spectral codes for each spatial pixel.
  • Fabry-Pérot Cavity: The FP cavity, formed by two distributed Bragg reflectors (DBRs) sandwiching the metasurface, acts as a resonant structure. The metasurface's phase shift shifts the resonant peak of the cavity, effectively creating a narrow-band spectral filter for each pixel.
  • Spatial-Spectral Encoding: When an object is placed near the SSE device, light from each spatial pixel of the object is passed through the corresponding spectral filter. This process maps the spatial information of the object onto a unique spectral signature for each pixel.
  • Computational Decoding: At the other end of the fiber, a spectrometer measures the encoded spectrum. A pseudo-inverse of the spectral encoding transfer matrix (M) is then used to computationally reconstruct the original spatial image from the measured spectrum. This matrix is characterized by measuring the spectral codes of individual pixels and the background.
  • Robustness Testing: The paper demonstrates the robustness of the system by testing image recovery with fiber bending and movement, and by evaluating the impact of spectral resolution on decoding accuracy.

Datasets

The paper does not explicitly mention pre-existing datasets for training or evaluation. Instead, it utilizes a custom-designed 4x4 binary mask to create test patterns. The performance is evaluated by comparing the recovered binary patterns to the original patterns from this mask.

Key Insights and Ideas

  • Metasurfaces as Spectral Filters: The innovative use of metasurfaces to create highly controllable and orthogonal spectral filters within an FP cavity is a key breakthrough. This allows for precise engineering of spectral codes, overcoming limitations of previous random or broad-band spectral encoding methods.
  • Non-Scanning Imaging: The ability to achieve imaging without any mechanical scanning is a significant advantage, leading to more compact, robust, and potentially faster imaging systems.
  • Robustness to Fiber Perturbations: The fact that the image remains largely unchanged with fiber bending or movement is crucial for practical applications in flexible endoscopic procedures. This robustness stems from the inherent nature of spectral encoding, which is less susceptible to physical distortions of the fiber compared to spatial encoding methods.
  • Low-Resolution Spectrometer Compatibility: The finding that decoding accuracy is only marginally affected by reduced spectral resolution suggests that the system can be implemented with miniaturized, low-cost spectrometers, further enhancing its practicality.

Key Results

  • The SSE device successfully captured 4x4 binary patterns with an average error rate of 9.8% when using a high-resolution spectrometer (0.2nm).
  • The error rate increased to approximately 11% even when the spectral resolution was reduced to 1.6nm, demonstrating the system's tolerance to lower-resolution measurements.
  • The recovered images remained largely unchanged when the fiber was bent or moved, highlighting the system's robustness.
  • The authors suggest that the number of pixels can be increased by using higher Q cavities or photonic crystal filters.

Limitations and Future Work

  • Pixel Misalignment: The authors attribute the ~10% average error rate to potential misalignment between the binary pattern on the mask and the SSE device.
  • Near-Field Imaging Assumption: The current SSE device is designed for near-field imaging. While the paper discusses using a lens to enable far-field imaging, this adds complexity.
  • Pixel Size Limitation: For smaller pixel sizes, satisfying the near-field condition becomes more challenging, potentially limiting applications in high-resolution microscopy without additional optics.
  • Future Work: The authors suggest increasing the number of pixels in the SSE device by improving the cavity's Q factor or using photonic crystal spectral filters. They also propose using a focusing lens to enable far-field imaging and potentially reduce pixel size to 2µm x 2µm. Further exploration of applications in compact endoscopes and angioscopes is also envisioned.

Summary updated

Abstract

Recent advancements in neuroimaging and microsurgery have sparked an increasing demand to capture images with miniaturized optical probes such as optical fibers. In this work, we present an approach to acquire images through a single fiber without the need for mechanical scanning. At the distal end of the fiber, a metasurface filter array encodes spatial information into a highly orthogonal spectrum. At the proximal end, the object can then be computationally recovered via the pseudo inverse of the encoding process. We demonstrate captures of a 4 × 4 binary object at the proximity of the spectral filter array using a 560–625 nm wavelength band. The recovered image maintains an error rate of <11% when measured using a spectrometer with a spectral resolution of 1.5 nm. Importantly, this modality remains unchanged as the fiber is bent or moved. Thus, our approach shows a robust way to image through a single optical fiber, with potential applications in compact endoscopes and angioscopes.
BibTeX
@inproceedings{xie2024spectrally,
  author = {Ningzhi Xie and Quentin A. A. Tanguy and Johannes E. Fröch and K. F. Böhringer and Arka Majumdar},
  title = {{Spectrally Encoded Nonscanning Imaging through a Fiber}},
  booktitle = {ACS},
  year = {2024},
  doi = {10.1021/acsphotonics.3c01582},
}
Figure