Deep spectral learning for label-free optical imaging oximetry with uncertainty quantification

Abstract

Measurement of blood oxygen saturation (sO2) by optical imaging oximetry provides invaluable insight into local tissue functions and metabolism. Despite different embodiments and modalities, all label-free optical imaging oximetry utilize the same principle of sO2-dependent spectral contrast from hemoglobin. Traditional approaches for quantifying sO2 often rely on analytical models that are fitted by the spectral measurements. These approaches in practice suffer from uncertainties due to biological variability, tissue geometry, light scattering, systemic spectral bias, and variations in experimental conditions. Here, we propose a new data-driven approach, termed deep spectral learning (DSL) for oximetry to be highly robust to experimental variations, and more importantly to provide uncertainty quantification for each sO2 prediction. To demonstrate the robustness and generalizability of DSL, we analyze data from two visible light optical coherence tomography (vis-OCT) setups across two separate in vivo experiments in rat retina. Predictions made by DSL are highly adaptive to experimental variabilities as well as the depth-dependent backscattering spectra. Two neural-network-based models are tested and compared with the traditional least-squares fitting (LSF) method. The DSL-predicted sO2 shows significantly lower mean-square errors than the LSF. For the first time, we have demonstrated en face maps of retinal oximetry along with pixel-wise confidence assessment. Our DSL overcomes several limitations in the traditional approaches and provides a more flexible, robust, and reliable deep learning approach for in vivo non-invasive label-free optical oximetry.

Single-cell cytometry via multiplexed fluorescence prediction by label-free reflectance microscopy

“…For many biomedical applications, it is beneficial to understand how much error the model may make without knowing the ground truth, i.e., the confidence of the model predictions. Emerging Bayesian DL-based uncertainty quantification techniques have proved useful to provide a proxy estimate of the prediction accuracy and quantify the model confidence (39,40), which will be adapted in our future work.…”

A review of uncertainty quantification in deep learning: Techniques, applications and challenges

“…Traditional methods for quantifying sO 2 suffer from uncertainties due to variations in the experimental conditions, systemic spectral bias, light spectral bias, tissue geometry and biological variability. Liu et al [430] devised deep spectral learning (DSL), a novel data-driven approach to yield oximetry that was robust to experimental variations and also facilitated uncertainty quantification for each sO 2 prediction. Predictions calculated by DSL were highly adaptive to the depth-dependent backscattering spectra as well as to experimental variabilities.…”

Uncertainty Quantification for Deep Unrolling-Based Computational Imaging

“…The problem of uncertainty quantification for image reconstruction tasks, e.g., [20]- [41], has attracted the attention of the computational imaging community again recently due to recent advancements in deep generative modeling [42] and BNNs [17]- [19]. The state-of-the-art deep learning-based image reconstruction methods performing uncertainty characterization, e.g., [20], [22], [25]- [33], [35]- [41], can be divided into two groups: deep generative model-based reconstruction methods and BNN-based reconstruction methods.…”