An official website of the United States government
Abstract Deep learning has been broadly applied to imaging in scattering applications. A common framework is to train a descattering network for image recovery by removing scattering artifacts. To achieve the best results on a broad spectrum of scattering conditions, individual “expert” networks need to be trained for each condition. However, the expert’s performance sharply degrades when the testing condition differs from the training. An alternative brute-force approach is to train a “generalist” network using data from diverse scattering conditions. It generally requires a larger network to encapsulate the diversity in the data and a sufficiently large training set to avoid overfitting. Here, we propose anadaptive learningframework, termed dynamic synthesis network (DSN), whichdynamicallyadjusts the model weights andadaptsto different scattering conditions. The adaptability is achieved by a novel “mixture of experts” architecture that enables dynamically synthesizing a network by blending multiple experts using a gating network. We demonstrate the DSN in holographic 3D particle imaging for a variety of scattering conditions. We show in simulation that our DSN provides generalization across acontinuumof scattering conditions. In addition, we show that by training the DSN entirely on simulated data, the network can generalize to experiments and achieve robust 3D descattering. We expect the same concept can find many other applications, such as denoising and imaging in scattering media. Broadly, our dynamic synthesis framework opens up a new paradigm for designing highlyadaptivedeep learning and computational imaging techniques.
more »
« less
Displacement-agnostic coherent imaging through scatter with an interpretable deep neural network
Coherent imaging through scatter is a challenging task. Both model-based and data-driven approaches have been explored to solve the inverse scattering problem. In our previous work, we have shown that a deep learning approach can make high-quality and highly generalizable predictions through unseen diffusers. Here, we propose a new deep neural network model that is agnostic to a broader class of perturbations including scatterer change, displacements, and system defocus up to 10× depth of field. In addition, we develop a new analysis framework for interpreting the mechanism of our deep learning model and visualizing its generalizability based on an unsupervised dimension reduction technique. We show that our model can unmix the scattering-specific information and extract the object-specific information and achieve generalization under different scattering conditions. Our work paves the way to arobustandinterpretabledeep learning approach to imaging through scattering media.
more »
« less
- Award ID(s):
- 1813848
- PAR ID:
- 10209809
- Publisher / Repository:
- Optical Society of America
- Date Published:
- Journal Name:
- Optics Express
- Volume:
- 29
- Issue:
- 2
- ISSN:
- 1094-4087; OPEXFF
- Format(s):
- Medium: X Size: Article No. 2244
- Size(s):
- Article No. 2244
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Laser speckle contrast imaging is widely used in clinical studies to monitor blood flow distribution. Speckle contrast tomography, similar to diffuse optical tomography, extends speckle contrast imaging to provide deep tissue blood flow information. However, the current speckle contrast tomography techniques suffer from poor spatial resolution and involve both computation and memory intensive reconstruction algorithms. In this work, we present SpeckleCam, a camera-based system to reconstruct high resolution 3D blood flow distribution deep inside the skin. Our approach replaces the traditional forward model using diffuse approximations with Monte-Carlo simulations-based convolutional forward model, which enables us to develop an improved deep tissue blood flow reconstruction algorithm. We show that our proposed approach can recover complex structures up to 6 mm deep inside a tissue-like scattering medium in the reflection geometry. We also conduct human experiments to demonstrate that our approach can detect reduced flow in major blood vessels during vascular occlusion.more » « less
-
Abstract BackgroundDue to intrinsic differences in data formatting, data structure, and underlying semantic information, the integration of imaging data with clinical data can be non‐trivial. Optimal integration requires robust data fusion, that is, the process of integrating multiple data sources to produce more useful information than captured by individual data sources. Here, we introduce the concept offusion qualityfor deep learning problems involving imaging and clinical data. We first provide a general theoretical framework and numerical validation of our technique. To demonstrate real‐world applicability, we then apply our technique to optimize the fusion of CT imaging and hepatic blood markers to estimate portal venous hypertension, which is linked to prognosis in patients with cirrhosis of the liver. PurposeTo develop a measurement method of optimal data fusion quality deep learning problems utilizing both imaging data and clinical data. MethodsOur approach is based on modeling the fully connected layer (FCL) of a convolutional neural network (CNN) as a potential function, whose distribution takes the form of the classical Gibbs measure. The features of the FCL are then modeled as random variables governed by state functions, which are interpreted as the different data sources to be fused. The probability density of each source, relative to the probability density of the FCL, represents a quantitative measure of source‐bias. To minimize this source‐bias and optimize CNN performance, we implement a vector‐growing encoding scheme called positional encoding, where low‐dimensional clinical data are transcribed into a rich feature space that complements high‐dimensional imaging features. We first provide a numerical validation of our approach based on simulated Gaussian processes. We then applied our approach to patient data, where we optimized the fusion of CT images with blood markers to predict portal venous hypertension in patients with cirrhosis of the liver. This patient study was based on a modified ResNet‐152 model that incorporates both images and blood markers as input. These two data sources were processed in parallel, fused into a single FCL, and optimized based on our fusion quality framework. ResultsNumerical validation of our approach confirmed that the probability density function of a fused feature space converges to a source‐specific probability density function when source data are improperly fused. Our numerical results demonstrate that this phenomenon can be quantified as a measure of fusion quality. On patient data, the fused model consisting of both imaging data and positionally encoded blood markers at the theoretically optimal fusion quality metric achieved an AUC of 0.74 and an accuracy of 0.71. This model was statistically better than the imaging‐only model (AUC = 0.60; accuracy = 0.62), the blood marker‐only model (AUC = 0.58; accuracy = 0.60), and a variety of purposely sub‐optimized fusion models (AUC = 0.61–0.70; accuracy = 0.58–0.69). ConclusionsWe introduced the concept of data fusion quality for multi‐source deep learning problems involving both imaging and clinical data. We provided a theoretical framework, numerical validation, and real‐world application in abdominal radiology. Our data suggests that CT imaging and hepatic blood markers provide complementary diagnostic information when appropriately fused.more » « less
-
Abstract We study the use of deep learning techniques to reconstruct the kinematics of the neutral current deep inelastic scattering (DIS) process in electron–proton collisions. In particular, we use simulated data from the ZEUS experiment at the HERA accelerator facility, and train deep neural networks to reconstruct the kinematic variables$$Q^2$$ andx. Our approach is based on the information used in the classical construction methods, the measurements of the scattered lepton, and the hadronic final state in the detector, but is enhanced through correlations and patterns revealed with the simulated data sets. We show that, with the appropriate selection of a training set, the neural networks sufficiently surpass all classical reconstruction methods on most of the kinematic range considered. Rapid access to large samples of simulated data and the ability of neural networks to effectively extract information from large data sets, both suggest that deep learning techniques to reconstruct DIS kinematics can serve as a rigorous method to combine and outperform the classical reconstruction methods.more » « less
-
Recovering 3D phase features of complex biological samples traditionally sacrifices computational efficiency and processing time for physical model accuracy and reconstruction quality. Here, we overcome this challenge using an approximant-guided deep learning framework in a high-speed intensity diffraction tomography system. Applying a physics model simulator-based learning strategy trained entirely on natural image datasets, we show our network can robustly reconstruct complex 3D biological samples. To achieve highly efficient training and prediction, we implement a lightweight 2D network structure that utilizes a multi-channel input for encoding the axial information. We demonstrate this framework on experimental measurements of weakly scattering epithelial buccal cells and strongly scatteringC. elegansworms. We benchmark the network’s performance against a state-of-the-art multiple-scattering model-based iterative reconstruction algorithm. We highlight the network’s robustness by reconstructing dynamic samples from a living worm video. We further emphasize the network’s generalization capabilities by recovering algae samples imaged from different experimental setups. To assess the prediction quality, we develop a quantitative evaluation metric to show that our predictions are consistent with both multiple-scattering physics and experimental measurements.more » « less
