We demonstrate highly accurate and fast cell imaging and classification systems enabled by the combination of disordered optical fiber for data transport and deep convolutional neural networks (DCNNs) for data analysis. Disordered optical fiber feature unique light transport properties based on the principle of transverse Anderson localization. A dense network of single mode-like transmission channels results in high spatial resolution while providing robustness regarding bending and environmental changes. DCNNs optimized for cell image reconstruction or cell classification have been trained and applied to perform rigorous testing. We show artifact-free real-time image reconstruction and >90% correct classification of cell samples.
As a label-free and quantitative imaging technique, optical diffraction tomography has been widely used in biological imaging. However, it is typically limited to weakly-scattering objects. To overcome this limitation, optimization algorithms based on minimizing field differences at the exit/observation plane, including total variation regularization, have been proposed and demonstrated. We propose a novel optimization algorithm to generalize field discrepancies from one plane to multiple planes throughout the scattering area. We numerically demonstrate that minimizing the field discrepancies at multiple planes instead of only one plane improves the robustness and accuracy of reconstructing multiply-scattering objects, without sacrificing the computational efficiency.
We demonstrate a deep-learning-based fiber imaging system that can transfer real-time artifact-free cell images through a meter-long Anderson localizing optical fiber. The cell samples are illuminated by an incoherent LED light source. A deep convolutional neural network is applied to the image reconstruction process. The network training uses data generated by a setup with straight fiber at room temperature (∼20 ° C) but can be utilized directly for high-fidelity reconstruction of cell images that are transported through fiber with a few degrees bend or fiber with segments heated up to 50°C. In addition, cell images located several millimeters away from the bare fiber end can be transported and recovered successfully without the assistance of distal optics. We provide evidence that the trained neural network is able to transfer its learning to recover images of cells featuring very different morphologies and classes that are never “seen” during the training process.
Computational imaging systems apply encoding on the physical layer of the imaging device, demonstrating superior performance in resolution, dynamic range, and acquisition speed, compared to conventional point-to-point mapping imaging system. However, accurate mathematical models is required for such systems, and the calibration is a major concern for practical implementation. In this invited talk, we will discuss the efforts in applying the learning approach in computational imaging system from the Optical Imaging System Lab at the University of Central Florida. Specifically, the talk will be focus on a demonstration of such approach in fully flexible lensless fiber imaging.
Transverse Anderson Localizing Optical Fiber (TALOF) provides a novel waveguiding mechanism. The entire disordered transverse structure of the fiber supports compactly located Anderson localized modes. Each transversely localized mode of the fiber forms a guiding channel. The strong disorder-induced transverse confinement due to the Anderson localization suffices for single-mode light transmission. In this work, we analyze the beam quality of highly localized modes in a glass-TALOF. We use the M^2 factor as a broadly accepted metric in optical fiber community for the beam quality evaluation. A numerical analysis of the M^2 values on a large ensemble of calculated modes in a glass-TALOF hints to the presence of high-quality modes at various locations across the transverse profile of the fiber. Our experimental results on the statistics of M^2 values in glass-TALOF supports the numerical analysis where it is shown that high-quality modes can be easily excited across the entire transverse profile of the fiber. Specifically, we present M^2 values of 30 localized modes and show how the M^2 ~ 1 modes are dominant. When the gain is added to TALOF, it can support localized lasing modes. Recently, we have demonstrated a directional and spectrally stable random laser mediated by a glass-TALOF; localization of the lasing modes reduces mode competition and helps with the spectral stability. The disorder-induced high-quality wavefronts in TALOF in combination with the aforementioned advances in lasing mediated by TALOF can foster a new class of single-mode optical fiber laser.