Zhichao Lin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhichao Lin has a total of 34 co-authors. Here are the top co-authors by number of shared publications:
- Maokun Li
Tsinghua University
16 shared publications - Shenheng Xu
Tsinghua University
15 shared publications - Fan Yang
Tsinghua University
14 shared publications - Rui Guo
Faculty of Math and Computer Science, Weizmann Institute of Science, Rehovot, Israel
13 shared publications - Aria Abubakar
SLB, Houston, TX, USA
11 shared publications - Ke Zhang
Department of Electronic Engineering, Beijing National Research Center for Information Science and Technology (BNRist), Tsinghua University, Beijing, China
4 shared publications
Latest Publications
- 25 October 2024
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 90  (3) , WA61-WA75
Deep joint inversion of multiple geophysical data with U-net reparameterization
Joint inversion integrates a variety of geophysical data in a unified framework to reveal different subsurface properties. Existing methods often require predefined correlations through mathematical equations or training data sets, heavily relying on prior information. In addition, integrating multiphysics data with different sensitivities is challenging. To address these limitations, we develop a novel approach that uses U-net to autonomously establish correlations during inversion, which yields consistent inverted models across different resolutions. Different properties are represented by U-net parameters and solved by minimizing multimodality data misfits simultaneously. We validate our approach through structure-constrained electromagnetic inversion, joint seismic and gravity inversion, and joint electromagnetic, seismic, and gravity inversion with synthetic data. This study affirms the feasibility of leveraging machine intelligence to autonomously integrate multiple geophysical data for precise subsurface characterization.
- 25 October 2024
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 90  (3) , WA61-WA75
Deep joint inversion of multiple geophysical data with U-net reparameterization
Joint inversion integrates a variety of geophysical data in a unified framework to reveal different subsurface properties. Existing methods often require predefined correlations through mathematical equations or training data sets, heavily relying on prior information. In addition, integrating multiphysics data with different sensitivities is challenging. To address these limitations, we develop a novel approach that uses U-net to autonomously establish correlations during inversion, which yields consistent inverted models across different resolutions. Different properties are represented by U-net parameters and solved by minimizing multimodality data misfits simultaneously. We validate our approach through structure-constrained electromagnetic inversion, joint seismic and gravity inversion, and joint electromagnetic, seismic, and gravity inversion with synthetic data. This study affirms the feasibility of leveraging machine intelligence to autonomously integrate multiple geophysical data for precise subsurface characterization.
- 26 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 72  (1) , 238-248
Attenuation Tomography Using Low-Frequency Ultrasound With Variational Autoencoder for Thorax Imaging: Experimental Study
In this paper, we introduce a novel inversion methodology employing the variational autoencoder (VAE) for human thorax attenuation tomography using low-frequency ultrasound. The VAE is trained to assimilate the structural priors of the human thorax, utilizing training samples generated from computed tomography (CT) scans. This approach enables the compression of high-dimensional attenuation distributions into a lowerdimensional latent space. During the inversion process, the latent code is optimized, and then the reconstructed model is generated by the decoder of the VAE. This process can effectively integrate prior information of the domain of interest (DOI) into the inversion through coding and decoding, which would mitigate the ill-posedness of the inverse problem and facilitate better outcomes. Our method demonstrates robust generalization capabilities and noise resilience in numerical simulations, outperforming the conventional pixel-based Gauss-Newton method. Human subject experiment further corroborates the effectiveness of our approach. This is also the first experimental validation of the feasibility of low-frequency ultrasound functional imaging of the human thorax. Although the current study presents certain limitations, it underscores the potential of low-frequency ultrasound in the continuous monitoring of the human respiratory system
- 26 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 72  (1) , 238-248
Attenuation Tomography Using Low-Frequency Ultrasound With Variational Autoencoder for Thorax Imaging: Experimental Study
In this paper, we introduce a novel inversion methodology employing the variational autoencoder (VAE) for human thorax attenuation tomography using low-frequency ultrasound. The VAE is trained to assimilate the structural priors of the human thorax, utilizing training samples generated from computed tomography (CT) scans. This approach enables the compression of high-dimensional attenuation distributions into a lowerdimensional latent space. During the inversion process, the latent code is optimized, and then the reconstructed model is generated by the decoder of the VAE. This process can effectively integrate prior information of the domain of interest (DOI) into the inversion through coding and decoding, which would mitigate the ill-posedness of the inverse problem and facilitate better outcomes. Our method demonstrates robust generalization capabilities and noise resilience in numerical simulations, outperforming the conventional pixel-based Gauss-Newton method. Human subject experiment further corroborates the effectiveness of our approach. This is also the first experimental validation of the feasibility of low-frequency ultrasound functional imaging of the human thorax. Although the current study presents certain limitations, it underscores the potential of low-frequency ultrasound in the continuous monitoring of the human respiratory system
- 28 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 43  (4) , 1365-1376
Three Dimensional Microwave Data Inversion in Feature Space for Stroke Imaging
Microwave imaging is a promising method for early diagnosing and monitoring brain strokes. It is portable, non-invasive, and safe to the human body. Conventional techniques solve for unknown electrical properties represented as pixels or voxels, but often result in inadequate structural information and high computational costs. We propose to reconstruct the three dimensional (3D) electrical properties of the human brain in a feature space, where the unknowns are latent codes of a variational autoencoder (VAE). The decoder of the VAE, with prior knowledge of the brain, acts as a module of data inversion. The codes in the feature space are optimized by minimizing the misfit between measured and simulated data. A dataset of 3D heads characterized by permittivity and conductivity is constructed to train the VAE. Numerical examples show that our method increases structural similarity by 14% and speeds up the solution process by over 3 orders of magnitude using only 4.8% number of the unknowns compared to the voxel-based method. This high-resolution imaging of electrical properties leads to more accurate stroke diagnosis and offers new insights into the study of the human brain.
- 28 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 43  (4) , 1365-1376
Three Dimensional Microwave Data Inversion in Feature Space for Stroke Imaging
Microwave imaging is a promising method for early diagnosing and monitoring brain strokes. It is portable, non-invasive, and safe to the human body. Conventional techniques solve for unknown electrical properties represented as pixels or voxels, but often result in inadequate structural information and high computational costs. We propose to reconstruct the three dimensional (3D) electrical properties of the human brain in a feature space, where the unknowns are latent codes of a variational autoencoder (VAE). The decoder of the VAE, with prior knowledge of the brain, acts as a module of data inversion. The codes in the feature space are optimized by minimizing the misfit between measured and simulated data. A dataset of 3D heads characterized by permittivity and conductivity is constructed to train the VAE. Numerical examples show that our method increases structural similarity by 14% and speeds up the solution process by over 3 orders of magnitude using only 4.8% number of the unknowns compared to the voxel-based method. This high-resolution imaging of electrical properties leads to more accurate stroke diagnosis and offers new insights into the study of the human brain.
- 30 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28  (1) , 367-378
Spatio-Temporal Classification of Lung Ventilation Patterns Using 3D EIT Images: A General Approach for Individualized Lung Function Evaluation
The Pulmonary Function Test (PFT) is a widely utilized and rigorous classification test for evaluating lung function, serving as a comprehensive diagnostic tool for lung conditions. Meanwhile, Electrical Impedance Tomography (EIT) is a rapidly advancing clinical technique that visualizes conductivity distribution induced by ventilation. EIT provides additional spatial and temporal information on lung ventilation beyond traditional PFT. However, relying solely on conventional isolated interpretations of PFT results and EIT images overlooks the continuous dynamic aspects of lung ventilation. This study aims to classify lung ventilation patterns by extracting spatial and temporal features from the 3D EIT image series. The study uses a Variational Autoencoder (VAE) with a MultiRes block to compress the spatial distribution in a 3D image into a one-dimensional vector. These vectors are then stacked to create a feature map for the exhibition of temporal features. A simple convolutional neural network is used for classification. Data from 137 subjects were utilized for the training phase. Initially, the model underwent validation through a leave-one-out cross-validation process. During this validation, the model achieved an accuracy and sensitivity of 0.96 and 1.00, respectively, with an f1-score of 0.98 when identifying the normal subjects. To assess pipeline reliability and feasibility, we tested it on 9 newly recruited subjects, with accurate ventilation mode predictions for 8 out of 9. In addition, we included 2D EIT results for comparison and conducted ablation experiments to validate the effectiveness of the VAE. The study demonstrates the potential of using image series for lung ventilation mode classification, providing a feasible method for patient prescreening and presenting an alternative form of PFT.
- 30 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 28  (1) , 367-378
Spatio-Temporal Classification of Lung Ventilation Patterns Using 3D EIT Images: A General Approach for Individualized Lung Function Evaluation
The Pulmonary Function Test (PFT) is a widely utilized and rigorous classification test for evaluating lung function, serving as a comprehensive diagnostic tool for lung conditions. Meanwhile, Electrical Impedance Tomography (EIT) is a rapidly advancing clinical technique that visualizes conductivity distribution induced by ventilation. EIT provides additional spatial and temporal information on lung ventilation beyond traditional PFT. However, relying solely on conventional isolated interpretations of PFT results and EIT images overlooks the continuous dynamic aspects of lung ventilation. This study aims to classify lung ventilation patterns by extracting spatial and temporal features from the 3D EIT image series. The study uses a Variational Autoencoder (VAE) with a MultiRes block to compress the spatial distribution in a 3D image into a one-dimensional vector. These vectors are then stacked to create a feature map for the exhibition of temporal features. A simple convolutional neural network is used for classification. Data from 137 subjects were utilized for the training phase. Initially, the model underwent validation through a leave-one-out cross-validation process. During this validation, the model achieved an accuracy and sensitivity of 0.96 and 1.00, respectively, with an f1-score of 0.98 when identifying the normal subjects. To assess pipeline reliability and feasibility, we tested it on 9 newly recruited subjects, with accurate ventilation mode predictions for 8 out of 9. In addition, we included 2D EIT results for comparison and conducted ablation experiments to validate the effectiveness of the VAE. The study demonstrates the potential of using image series for lung ventilation mode classification, providing a feasible method for patient prescreening and presenting an alternative form of PFT.
- 25 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72  (00189456) , 1-12
Tri-Modal Joint Inversion Based on Disentangled Variational Autoencoder for Human Thorax Imaging
In this work, we present a joint inversion framework that enables simultaneous reconstruction of electrical impedance tomography (EIT), microwave tomography (MWT), and ultra-sound tomography (UST). The structural and value features of the three modality images are decoupled and encoded into structure code and value code by the proposed disentangled variational autoencoder (DVAE). The inversion is then performed at the feature level, with the structure code and value code treated as unknown parameters to be inverted by the Gauss-Newton method. The interaction among the three modalities is achieved through structure code fusion, where the structure codes of the three modalities are averaged in the inversion as the fused structure code, thus ensuring the consistency of the structures in the three modalities. Two schemes are proposed to utilize the fused structure code. The initial code scheme (ICS) employs it as the initial structure code for the Gauss-Newton update of each modality. Meanwhile, the reference code scheme (RCS) designates it as the reference structure code. Numerical examples of human thorax imaging verify the effectiveness of the two schemes and demonstrate the advantages of the proposed joint inversion framework over separate inversions.
- 25 September 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72  (00189456) , 1-12
Tri-Modal Joint Inversion Based on Disentangled Variational Autoencoder for Human Thorax Imaging
In this work, we present a joint inversion framework that enables simultaneous reconstruction of electrical impedance tomography (EIT), microwave tomography (MWT), and ultra-sound tomography (UST). The structural and value features of the three modality images are decoupled and encoded into structure code and value code by the proposed disentangled variational autoencoder (DVAE). The inversion is then performed at the feature level, with the structure code and value code treated as unknown parameters to be inverted by the Gauss-Newton method. The interaction among the three modalities is achieved through structure code fusion, where the structure codes of the three modalities are averaged in the inversion as the fused structure code, thus ensuring the consistency of the structures in the three modalities. Two schemes are proposed to utilize the fused structure code. The initial code scheme (ICS) employs it as the initial structure code for the Gauss-Newton update of each modality. Meanwhile, the reference code scheme (RCS) designates it as the reference structure code. Numerical examples of human thorax imaging verify the effectiveness of the two schemes and demonstrate the advantages of the proposed joint inversion framework over separate inversions.
- 9 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 71  (4) , 3746-3751
Physics-Informed Supervised Residual Learning for 2-D Inverse Scattering Problems
In this paper, we propose a new physics-constrained approach to solve 2D inverse scattering problems (ISPs) by extending physics-informed supervised residual learning (PhiSRL) with Born approximation (BA). By embedding the fixed-point iteration method in residual neural network (ResNet), PhiSRL aims to solve ISPs iteratively by applying the convolutional neural networks (CNNs) to learn the update rules of reconstructions. PhiSRL is employed to invert lossy scatterers by introducing BA to linearize ISPs and further reduce the computational burden of forward modeling. Both numerical and experimental results validate the effectiveness of the proposed approach.
- 9 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 71  (4) , 3746-3751
Physics-Informed Supervised Residual Learning for 2-D Inverse Scattering Problems
In this paper, we propose a new physics-constrained approach to solve 2D inverse scattering problems (ISPs) by extending physics-informed supervised residual learning (PhiSRL) with Born approximation (BA). By embedding the fixed-point iteration method in residual neural network (ResNet), PhiSRL aims to solve ISPs iteratively by applying the convolutional neural networks (CNNs) to learn the update rules of reconstructions. PhiSRL is employed to invert lossy scatterers by introducing BA to linearize ISPs and further reduce the computational burden of forward modeling. Both numerical and experimental results validate the effectiveness of the proposed approach.
- 8 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72  (00189456) , 1-12
A Deep Generative Model-Integrated Framework for 3-D Time-Difference Electrical Impedance Tomography
The time-difference image reconstruction problem of electrical impedance tomography (EIT) refers to reconstructing the conductivity change in a human body part between two time points using the boundary impedance measurements. Conventionally, the problem can be formulated as a linear inverse problem. However, due to the physical property of the forward process, the inverse problem is seriously ill-posed. As a result, traditional regularized least-squares-based methods usually produce low-resolution images that are difficult to interpret. This work proposes a framework that uses a deep generative model to constrain the unknown conductivity. Specifically, this framework allows the inclusion of a constraint that describes a mathematical relationship between the generative model and the unknown conductivity. The resultant constrained minimization problem is solved using an extended alternating direction method of multipliers (ADMM). The effectiveness of the framework is demonstrated by the example of three-dimensional time-difference chest EIT imaging. Numerical experiment shows a significant improvement of the image quality compared with total variation-regularized least-squares method (PSNR is improved by 4.3% for 10% noise and 4.6% for 30% noise; SSIM is improved by 4.8% for 10% noise and 6.0% for 30% noise). Human experiments show improved correlation between the reconstructed images and images from reference techniques.
- 8 December 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72  (00189456) , 1-12
A Deep Generative Model-Integrated Framework for 3-D Time-Difference Electrical Impedance Tomography
The time-difference image reconstruction problem of electrical impedance tomography (EIT) refers to reconstructing the conductivity change in a human body part between two time points using the boundary impedance measurements. Conventionally, the problem can be formulated as a linear inverse problem. However, due to the physical property of the forward process, the inverse problem is seriously ill-posed. As a result, traditional regularized least-squares-based methods usually produce low-resolution images that are difficult to interpret. This work proposes a framework that uses a deep generative model to constrain the unknown conductivity. Specifically, this framework allows the inclusion of a constraint that describes a mathematical relationship between the generative model and the unknown conductivity. The resultant constrained minimization problem is solved using an extended alternating direction method of multipliers (ADMM). The effectiveness of the framework is demonstrated by the example of three-dimensional time-difference chest EIT imaging. Numerical experiment shows a significant improvement of the image quality compared with total variation-regularized least-squares method (PSNR is improved by 4.3% for 10% noise and 4.6% for 30% noise; SSIM is improved by 4.8% for 10% noise and 6.0% for 30% noise). Human experiments show improved correlation between the reconstructed images and images from reference techniques.
- 1 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 71  (1) , 818-829
Neural Born Iterative Method for Solving Inverse Scattering Problems: 2D Cases
In this paper, we propose the neural Born iterative method (NeuralBIM) for solving 2D inverse scattering problems (ISPs) by drawing on the scheme of physics-informed supervised residual learning (PhiSRL) to emulate the computing process of the traditional Born iterative method (TBIM). NeuralBIM employs independent convolutional neural networks (CNNs) to learn the alternate update rules of two different candidate solutions regarding the residuals. Two different schemes are presented in this paper, including the supervised and unsupervised learning schemes. With the data set generated by the method of moments (MoM), supervised NeuralBIM are trained with the knowledge of total fields and contrasts. Unsupervised NeuralBIM is guided by the physics-embedded objective function founding on the governing equations of ISPs, which results in no requirement of total fields and contrasts for training. Numerical and experimental results further validate the efficacy of NeuralBIM.
- 1 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 71  (1) , 818-829
Neural Born Iterative Method for Solving Inverse Scattering Problems: 2D Cases
In this paper, we propose the neural Born iterative method (NeuralBIM) for solving 2D inverse scattering problems (ISPs) by drawing on the scheme of physics-informed supervised residual learning (PhiSRL) to emulate the computing process of the traditional Born iterative method (TBIM). NeuralBIM employs independent convolutional neural networks (CNNs) to learn the alternate update rules of two different candidate solutions regarding the residuals. Two different schemes are presented in this paper, including the supervised and unsupervised learning schemes. With the data set generated by the method of moments (MoM), supervised NeuralBIM are trained with the knowledge of total fields and contrasts. Unsupervised NeuralBIM is guided by the physics-embedded objective function founding on the governing equations of ISPs, which results in no requirement of total fields and contrasts for training. Numerical and experimental results further validate the efficacy of NeuralBIM.
- 8 August 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 70  (11) , 11059-11069
A Nonlinear Model Compression Scheme Based on Variational Autoencoder for Microwave Data Inversion
We present an inversion algorithm with a deep-learning-based model compression scheme. Models are described with latent parameters of a trained variational autoencoder (VAE) neural network. Given observed data, latent parameters are inverted by minimizing the data misfit cost function using the Gauss-Newton method. This inversion algorithm is tested using both synthetic and experimental datasets. We achieve a 0.87% compression rate while maintaining high-quality reconstruction. The deep neural network renders nonlinear model compression, which largely reduces the number of unknowns; hence, it has higher computational efficiency. Furthermore, various prior knowledge that is difficult to describe with rigorous forms can be incorporated into inversion through training the neural network, which mitigates the ill-posedness of the inverse problem.
- 8 August 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Antennas and Propagation
- Vol. 70  (11) , 11059-11069
A Nonlinear Model Compression Scheme Based on Variational Autoencoder for Microwave Data Inversion
We present an inversion algorithm with a deep-learning-based model compression scheme. Models are described with latent parameters of a trained variational autoencoder (VAE) neural network. Given observed data, latent parameters are inverted by minimizing the data misfit cost function using the Gauss-Newton method. This inversion algorithm is tested using both synthetic and experimental datasets. We achieve a 0.87% compression rate while maintaining high-quality reconstruction. The deep neural network renders nonlinear model compression, which largely reduces the number of unknowns; hence, it has higher computational efficiency. Furthermore, various prior knowledge that is difficult to describe with rigorous forms can be incorporated into inversion through training the neural network, which mitigates the ill-posedness of the inverse problem.
- 18 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 71  (00189456) , 1-12
Feature-Based Inversion Using Variational Autoencoder for Electrical Impedance Tomography
A feature-based inversion method is presented to incorporate structural prior information of the human thorax for absolute imaging in electrical impedance tomography (EIT). A set of EIT images are generated from open-source computed tomography (CT) scans with embedded structural priors of the human thorax. A variational autoencoder (VAE) is applied to learn high-level features of these EIT images and construct a mapping between EIT images and latent codes in a low-dimensional feature space. Then, the parameters of the latent code are served as unknowns to be inverted under a deterministic framework by the Gauss–Newton (GN) method. In this way, the number of unknowns is greatly reduced. Both synthetic and experimental data validate the proposed method. The reconstructed image quality is significantly improved, and the proposed method is relatively robust to measurement noise and modeling errors. This method provides a flexible and effective way to incorporate structural information from CT scans into EIT inversion, which is a potential algorithm for human thorax imaging.
- 18 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 71  (00189456) , 1-12
Feature-Based Inversion Using Variational Autoencoder for Electrical Impedance Tomography
A feature-based inversion method is presented to incorporate structural prior information of the human thorax for absolute imaging in electrical impedance tomography (EIT). A set of EIT images are generated from open-source computed tomography (CT) scans with embedded structural priors of the human thorax. A variational autoencoder (VAE) is applied to learn high-level features of these EIT images and construct a mapping between EIT images and latent codes in a low-dimensional feature space. Then, the parameters of the latent code are served as unknowns to be inverted under a deterministic framework by the Gauss–Newton (GN) method. In this way, the number of unknowns is greatly reduced. Both synthetic and experimental data validate the proposed method. The reconstructed image quality is significantly improved, and the proposed method is relatively robust to measurement noise and modeling errors. This method provides a flexible and effective way to incorporate structural information from CT scans into EIT inversion, which is a potential algorithm for human thorax imaging.