Reconstruction for limited-data nonlinear tomographic absorption spectroscopy via deep learning

Huang, Jianqing; Liu, Hecong; Dai, Jinghang; Cai, Weiwei

doi:10.1016/j.jqsrt.2018.07.011

Cited by 62 publications

(27 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Convolutional neural nets (CNNs) have been particularly successful in the realm of image processing and were therefore adapted for tomographic reconstruction by numerous groups. For instance, Huang et al [27], [28], [29] conducted 2D laser absorption tomography and 3D chemiluminescence tomography via CNNs, and Wei et al [30], [31] used CNNs to reconstruct 3D mole fraction and temperature fields from laser absorption measurements of methane and ethylene flame doublets. These examples employed the common supervised training paradigm, in which phantom distributions (c exact ) paired with synthetic measurements (b exact = Ac exact ) are used to train the network.…”

Section: B Deep Learning Reconstruction Algorithmsmentioning

confidence: 99%

“…These examples employed the common supervised training paradigm, in which phantom distributions (c exact ) paired with synthetic measurements (b exact = Ac exact ) are used to train the network. Phantoms for training have been obtained from previous reconstructions [28], [29], random Gaussian fields [27], [30], and large-eddy simulations [31], and random errors were usually added to b train to make the reconstruction procedure robust to noise. While this approach can yield accurate estimates of the QoI when the training set accurately represents the target physics, the application of DNN-based reconstruction to targets that exhibit unique structures is not reliable.…”

Section: B Deep Learning Reconstruction Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Flow Field Tomography with Uncertainty Quantification using a Bayesian Physics-Informed Neural Network

Molnar,

Grauer

2021

Preprint

View full text Add to dashboard Cite

We report a new approach to flow field tomography that uses the Navier-Stokes and advection-diffusion equations to regularize reconstructions. Tomography is increasingly employed to infer 2D or 3D fluid flow and combustion structures from a series of lineof-sight (LoS) integrated measurements using a wide array of imaging modalities. The high-dimensional flow field is reconstructed from low-dimensional measurements by inverting a projection model that comprises path integrals along each LoS through the region of interest. Regularization techniques are needed to obtain realistic estimates, but current methods rely on a penalty term that is incompatible with the flow physics to varying degrees. Physics-informed neural networks (PINNs) provide a new framework for inverse analysis that enables regularization of the flow field estimates using the governing physics. We demonstrate how a PINN can be leveraged to reconstruct a 2D flow field from sparse LoS-integrated measurements with no knowledge of the boundary conditions by incorporating the measurement model into the loss function used to train the network. The resulting reconstructions are remarkably superior to reconstructions produced by state-of-the-art algorithms, even when a PINN is used for post-processing. However, as with conventional iterative algorithms, our approach is susceptible to semi-convergence when there is a high level of noise. We address this issue through the use of a Bayesian PINN, which facilitates comprehensive uncertainty quantification of the reconstructions, enables the use of a more intuitive loss function, and reveals the source of semi-convergence.

show abstract

Section: B Deep Learning Reconstruction Algorithmsmentioning

confidence: 99%

Section: B Deep Learning Reconstruction Algorithmsmentioning

confidence: 99%

Flow Field Tomography with Uncertainty Quantification using a Bayesian Physics-Informed Neural Network

Molnar,

Grauer

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…However, excessive filtering will lead to the loss of some helpful information in the ECG signal. Since CNN has better noise immunity (Huang et al, 2018), we only perform baseline wandering removal and preserve as much information as possible from the raw ECG signal.…”

Section: Ecg Denoisingmentioning

confidence: 99%

A hybrid method for heartbeat classification via convolutional neural networks, multilayer perceptrons and focal loss

Wang

Yang³

et al. 2020

PeerJ Computer Science

View full text Add to dashboard Cite

Background Heart arrhythmia, as one of the most important cardiovascular diseases (CVDs), has gained wide attention in the past two decades. The article proposes a hybrid method for heartbeat classification via convolutional neural networks, multilayer perceptrons and focal loss. Methods In the method, a convolution neural network is used to extract the morphological features. The reason behind this is that the morphological characteristics of patients have inter-patient variations, which makes it difficult to accurately describe using traditional hand-craft ways. Then the extracted morphological features are combined with the RR intervals features and input into the multilayer perceptron for heartbeat classification. The RR intervals features contain the dynamic information of the heartbeat. Furthermore, considering that the heartbeat classes are imbalanced and would lead to the poor performance of minority classes, a focal loss is introduced to resolve the problem in the article. Results Tested using the MIT-BIH arrhythmia database, our method achieves an overall positive predictive value of 64.68%, sensitivity of 68.55%, f1-score of 66.09%, and accuracy of 96.27%. Compared with existing works, our method significantly improves the performance of heartbeat classification. Conclusions Our method is simple yet effective, which is potentially used for personal automatic heartbeat classification in remote medical monitoring. The source code is provided on https://github.com/JackAndCole/Deep-Neural-Network-For-Heartbeat-Classification.

show abstract

“…CNN has been demonstrated recently in CST simulations to perform spatially resolved measurements in combustion diagnosis [17], [18]. Previous work employed CNNs in CST and showed that their models could achieve a similar accuracy level as simulated annealing [17] and the reduction in network parameters [18].…”

Section: Introductionmentioning

confidence: 99%

CSTNet: A Dual-Branch Convolutional Neural Network for Imaging of Reactive Flows Using Chemical Species Tomography

Jiang

Zhang

et al. 2023

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Chemical species tomography (CST) has been widely used for in situ imaging of critical parameters, e.g., species concentration and temperature, in reactive flows. However, even with state-of-the-art computational algorithms, the method is limited due to the inherently ill-posed and rankdeficient tomographic data inversion and by high computational cost. These issues hinder its application for real-time flow diagnosis. To address them, we present here a novel convolutional neural network, namely CSTNet, for high-fidelity, rapid, and simultaneous imaging of species concentration and temperature using CST. CSTNet introduces a shared feature extractor that incorporates the CST measurements and sensor layout into the learning network. In addition, a dual-branch decoder with internal crosstalk, which automatically learns the naturally correlated distributions of species concentration and temperature, is proposed for image reconstructions. The proposed CSTNet is validated both with simulated datasets and with measured data from real flames in experiments using an industry-oriented sensor. Superior performance is found relative to previous approaches in terms of reconstruction accuracy and robustness to measurement noise. This is the first time, to the best of our knowledge, that a deep learning-based method for CST has been experimentally validated for simultaneous imaging of multiple critical parameters in reactive flows using a low-complexity optical sensor with a severely limited number of laser beams.

show abstract

Reconstruction for limited-data nonlinear tomographic absorption spectroscopy via deep learning

Cited by 62 publications

References 30 publications

Flow Field Tomography with Uncertainty Quantification using a Bayesian Physics-Informed Neural Network

Flow Field Tomography with Uncertainty Quantification using a Bayesian Physics-Informed Neural Network

A hybrid method for heartbeat classification via convolutional neural networks, multilayer perceptrons and focal loss

CSTNet: A Dual-Branch Convolutional Neural Network for Imaging of Reactive Flows Using Chemical Species Tomography

Contact Info

Product

Resources

About