Hybrid iterating-averaging low photon budget Gabor holographic microscopy (2409.14746v1)

Abstract: One of the primary challenges in live cell culture observation is achieving high-contrast imaging with minimal impact on sample behavior. Quantitative phase imaging (QPI) techniques address this by providing label-free, high-contrast images of transparent samples. Measurement system influence may be further reduced by imaging samples under low illumination intensity (low photon budget - LPB), thereby minimizing photostimulation, phototoxicity, and photodamage, and enabling high-speed imaging. LPB imaging is challenging in QPI due to significant camera shot noise and quantification noise. Digital in-line holographic microscopy (DIHM), working with or without lenses, is a QPI technique known for its robustness to quantification noise. However, reducing simultaneously the camera shot noise and the inherent in-line holographic twin image disturbances remain a critical, yet unaddressed, challenge. This study introduces an innovative iterative Gabor averaging (IGA) algorithm designed specifically for filling this important scientific gap in multi-frame DIHM under LPB conditions. We evaluated the performance of the IGA on simulated data showing that it outperformed traditional algorithms in terms of reconstruction accuracy under high noise conditions. Those results were corroborated by experimental validation involving high-speed imaging of dynamic sperm cells and a phase test target under significantly reduced illumination power. Additionally, the IGA algorithm proved successful in reconstructing optically thin samples, which typically produce low signal-to-noise ratio holograms even under high photon budget conditions. These advancements facilitate photostimulation-free and high-speed imaging of dynamic biological samples and enhance the capability to image samples with extremely low optical thickness, potentially transforming various applications in biomedical and environmental imaging.

• The paper introduces an iterative Gabor averaging (IGA) algorithm that effectively minimizes shot noise and twin image interference in low photon budget DIHM.
• Simulations and high-speed experiments confirmed that the IGA method improves phase reconstruction accuracy and maintains high spatial resolution under challenging imaging conditions.
• The study’s approach enables non-invasive, high-speed imaging of dynamic biological samples, paving the way for advanced applications in quantitative phase imaging.

Hybrid Iterating-Averaging Low Photon Budget Gabor Holographic Microscopy

The paper "Hybrid Iterating-Averaging Low Photon Budget Gabor Holographic Microscopy" by Mikołaj Rogalski et al. presents innovative advancements in quantitative phase imaging (QPI) aimed at addressing significant challenges in live cell culture observation. Specifically, the paper introduces a novel iterative Gabor averaging (IGA) algorithm designed to enhance digital in-line holographic microscopy (DIHM) performance under low photon budget (LPB) conditions.

Context and Problem Statement

Observing live bio-samples such as cell cultures through optical microscopy is notably challenging due to the inherent transparency of these samples, which typically results in very low contrast images. QPI techniques, and DIHM in particular, have emerged as promising solutions for visualizing transparent objects with high contrast by measuring the phase delay between light waves passing through the sample versus its surrounding medium. While DIHM is appreciated for its simplicity, cost-effectiveness, and robustness against quantification noise, it still struggles with significant camera shot noise and the twin image disturbance, especially under LPB conditions.

Contributions

The key contribution of this paper is the development and validation of the IGA algorithm, which effectively addresses the simultaneous minimization of shot noise and twin image disturbance in LPB DIHM. The algorithm combines iterative phase retrieval techniques and frame averaging, leveraging the advantages of both approaches while mitigating their respective drawbacks.

Methodology

1. Digital In-Line Holographic Microscopy Systems:
   • Two DIHM setups were employed: a lens-based system configured with a conventional bright-field microscope and a lensless setup designed for higher field-of-view with minimal hardware.
   • Both systems utilize in-line holography principles, where the camera captures the amplitude part of the free-space propagated optical field, which can then be numerically reconstructed to retrieve phase information.
2. Iterative Gabor Averaging Algorithm:
   • The IGA method integrates iterative phase retrieval with Gabor averaging, where the former focuses on minimizing low-frequency twin image disturbances and the latter addresses high-frequency shot noise.
   • The algorithm numerically processes multiple in-line holograms collected under varied imaging conditions, enhancing the reconstruction accuracy across different noise levels.

Results

The effectiveness of the IGA algorithm was demonstrated through both simulations and experimental validations:

• Simulations:
  • The algorithm showed marked improvements in phase reconstruction accuracy over traditional algorithms (Gerchberg-Saxton (GS) and Gabor averaging (GA)), especially under high shot noise conditions.
  • The IGA algorithm's ability to maintain resolution while minimizing noise was robust across different simulated noise levels.
• Experimental Validation:
  • High-speed imaging of dynamic sperm cells showcased the IGA method's capability to produce clear images under extremely low illumination conditions.
  • The reconstructed images from optically thin phase test targets further affirmed the algorithm’s prowess in maintaining spatial resolution and accurately reconstructing low SNR holographic data.

Implications

The proposed IGA algorithm offers substantial advancements in QPI, markedly enhancing DIHM's adaptability and reliability under LPB conditions. These improvements can lead to significant benefits in practical applications such as high-speed imaging of living biological samples without causing photostimulation or photodamage, and imaging of samples with very low optical thickness.

Future Directions

Future research can explore further optimization and automatization of IGA parameters tailored to specific imaging scenarios. Additionally, the algorithm's framework can be extended to other computational imaging systems such as ptychography and Fourier ptychographic microscopy, which also employ iterative phase retrieval procedures.

Conclusion

The hybrid iterating-averaging approach embodied by the IGA algorithm presents a significant enhancement for DIHM under LPB conditions, addressing a critical gap in current imaging techniques. This research paves the way for more effective, non-invasive, and high-resolution imaging of dynamic and optically challenging biological specimens.

Funded by various prominent institutions and supported by a wealth of experimental and numerical data, this paper makes a substantial contribution to the field of biomedical imaging technology, promising to offer new tools and methods for researchers and practitioners engaged in live cell imaging and other related areas.