An official website of the United States government
Ultrasound computed tomography (USCT) is one of the advanced imaging techniques used in structural health monitoring (SHM) and medical imaging due to its relatively low-cost, rapid data acquisition process. The time-domain full waveform inversion (TDFWI) method, an iterative inversion approach, has shown great promise in USCT. However, such an iterative process can be very time-consuming and computationally expensive but can be greatly accelerated by integrating an AI-based approach (e.g., convolution neural network (CNN)). Once trained, the CNN model takes low-iteration TDFWI images as input and instantaneously predicts material property distribution within the scanned region. Nevertheless, the quality of the reconstruction with the current CNN degrades with the increased complexity of material distributions. Another challenge is the availability of enough experimental data and, in some cases, even synthetic surrogate data. To alleviate these issues, this paper details a systematic study of the enhancement effect of a 2D CNN (U-Net) by improving the quality with limited training data. To achieve this, different augmentation schemes (flipping and mixing existing data) were implemented to increase the amount and complexity of the training datasets without generating a substantial number of samples. The objective was to evaluate the enhancement effect of these augmentation techniques on the performance of the U-Net model at FWI iterations. A thousand numerically built samples with acoustic material properties are used to construct multiple datasets from different FWI iterations. A parallelized, high-performance computing (HPC) based framework has been created to rapidly generate the training data. The prediction results were compared against the ground truth images using standard matrices, such as the structural similarity index measure (SSIM) and average mean square error (MSE). The results show that the increased number of samples from augmentations improves shape imaging of the complex regions even with a low iteration FWI training data.
more »
« less
Learn low-wavenumber information in FWI via deep inception-based convolutional networks
In this paper, we will explore the possibility of synthesizing the low-frequency data from the high-frequency data. The synthesized low-frequency data are used to improve the full-waveform inversion (FWI). Unlike all previously methods, to the best of our knowledge, this is the first attempt to utilize a data driven approach to solve the problem. We propose to learn the low wavenumber information in FWI via the Deep Inception based Convolutional Networks. Once the deep learning network is sufficiently trained, the network can be used to predicted the low-frequency data with high accuracy on a completely different unknown velocity model. In the end, we validate the quality of the predicted low-frequency data and the robustness of this deep learning approach.
more »
« less
- Award ID(s):
- 1746824
- PAR ID:
- 10101080
- Date Published:
- Journal Name:
- SEG Technical Program Expanded Abstracts 2018
- Page Range / eLocation ID:
- 2091 to 2095
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Objective. Arterial viscosity is emerging as an important biomarker, in addition to the widely used arterial elasticity. This paper presents an approach to estimate arterial viscoelasticity using shear wave elastography (SWE).Approach. While dispersion characteristics are often used to estimate elasticity from SWE data, they are not sufficiently sensitive to viscosity. Driven by this, we develop a full waveform inversion (FWI) methodology, based on directly matching predicted and measured wall velocity in space and time, to simultaneously estimate both elasticity and viscosity. Specifically, we propose to minimize an objective function capturing the correlation between measured and predicted responses of the anterior wall of the artery.Results. The objective function is shown to be well-behaving (generally convex), leading us to effectively use gradient optimization to invert for both elasticity and viscosity. The resulting methodology is verified with synthetic data polluted with noise, leading to the conclusion that the proposed FWI is effective in estimating arterial viscoelasticity.Significance. Accurate estimation of arterial viscoelasticity, not just elasticity, provides a more precise characterization of arterial mechanical properties, potentially leading to a better indicator of arterial health.more » « less
-
Deep neural networks (DNNs) have showcased their remarkable precision in approximating smooth functions. However, they suffer from the spectral bias, wherein DNNs typically exhibit a tendency to prioritize the learning of lower-frequency components of a function, struggling to effectively capture its high-frequency features. This paper is to address this issue. Notice that a function having only low frequency components may be well-represented by a shallow neural network (SNN), a network having only a few layers. By observing that composition of low frequency functions can effectively approximate a high-frequency function, we propose to learn a function containing high-frequency components by composing several SNNs, each of which learns certain low-frequency information from the given data. We implement the proposed idea by exploiting the multi-grade deep learning (MGDL) model, a recently introduced model that trains a DNN incrementally, grade by grade, a current grade learning from the residue of the previous grade only an SNN (with trainable parameters) composed with the SNNs (with fixed parameters) trained in the preceding grades as features. We apply MGDL to synthetic, manifold, colored images, and MNIST datasets, all characterized by presence of high-frequency features. Our study reveals that MGDL excels at representing functions containing high-frequency information. Specifically, the neural networks learned in each grade adeptly capture some low-frequency information, allowing their compositions with SNNs learned in the previous grades effectively representing the high-frequency features. Our experimental results underscore the efficacy of MGDL in addressing the spectral bias inherent in DNNs. By leveraging MGDL, we offer insights into overcoming spectral bias limitation of DNNs, thereby enhancing the performance and applicability of deep learning models in tasks requiring the representation of high-frequency information. This study confirms that the proposed method offers a promising solution to address the spectral bias of DNNs. The code is available on GitHub: Addressing Spectral Bias via MGDL.more » « less
-
Abstract Ultrasound computed tomography (USCT) shows great promise in nondestructive evaluation and medical imaging due to its ability to quickly scan and collect data from a region of interest. However, existing approaches are a tradeoff between the accuracy of the prediction and the speed at which the data can be analyzed, and processing the collected data into a meaningful image requires both time and computational resources. We propose to develop convolutional neural networks (CNNs) to accelerate and enhance the inversion results to reveal underlying structures or abnormalities that may be located within the region of interest. For training, the ultrasonic signals were first processed using the full waveform inversion (FWI) technique for only a single iteration; the resulting image and the corresponding true model were used as the input and output, respectively. The proposed machine learning approach is based on implementing two-dimensional CNNs to find an approximate solution to the inverse problem of a partial differential equation-based model reconstruction. To alleviate the time-consuming and computationally intensive data generation process, a high-performance computing-based framework has been developed to generate the training data in parallel. At the inference stage, the acquired signals will be first processed by FWI for a single iteration; then the resulting image will be processed by a pre-trained CNN to instantaneously generate the final output image. The results showed that once trained, the CNNs can quickly generate the predicted wave speed distributions with significantly enhanced speed and accuracy.more » « less
-
We present a methodology based on the implementation of a fully connected neural network algorithm to estimate the temporal evolution of the high-frequency gravitational wave emission for a core collapse supernova (CCSN). For this study, we selected a fully connected deep neural network (DNN) regression model because it can learn both linear and nonlinear relationships between the input and output data, it is more appropriate for handling large-dimensional input data, and it offers high performance at a low computational cost. To train the Machine Learning (ML) algorithm, we construct a training dataset using synthetic waveforms, and several CCSN waveforms are used to test the algorithm. We performed a first-order estimation of the high-frequency gravitational wave emission on real interferometric LIGO data from the second half of the third observing run (O3b) with a two detector network (L1 and H1). The relative error associated with the estimate of the slope of the resonant frequency versus time for the GW from CCSN signals is within 13% for the tested candidates included in this study up to different Galactic distances (1.0, 2.3, 3.1, 4.3, 5.4, 7.3, and 10 kpc). This method is, to date, the best estimate of the temporal evolution of the high-frequency emission in real interferometric data. Our methodology of estimation can be used in future studies focused on physical properties of the progenitor. The distances where comparable performances could be achieved for Einstein Telescope and Cosmic Explorer roughly rescale with the noise floor improvements.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1190/segam2018-2997901.1
