Peng Hu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Peng Hu has a total of 72 co-authors. Here are the top co-authors by number of shared publications:
- Lihong Victory Wang
Caltech Optical Imaging Laboratory, Andrew and Peggy Cherng Department of Medical Engineering, Department of Electrical Engineering, California Institute of Technology, Pasadena, CA 91125, USA
15 shared publications Li Lin
Caltech Optical Imaging Laboratory, Andrew and Peggy Cherng Department of Medical Engineering, California Institute of Technology, Pasadena, USA
6 shared publicationsXin Tong
Caltech Optical Imaging Laboratory, Andrew and Peggy Cherng Department of Medical Engineering, California Institute of Technology, Pasadena, USA
6 shared publicationsKonstantin Maslov
Caltech Optical Imaging Laboratory, Andrew and Peggy Cherng Department of Medical Engineering, California Institute of Technology, Pasadena, USA
4 shared publicationsJunhui Shi
Caltech Optical Imaging Laboratory, Andrew and Peggy Cherng Department of Medical Engineering, California Institute of Technology, Pasadena, USA
3 shared publicationsLei Li
Caltech Optical Imaging Laboratory, Andrew and Peggy Cherng Department of Medical Engineering, Department of Electrical Engineering, California Institute of Technology, Pasadena, USA
3 shared publications
Latest Publications
- Top Cited Papers
- 16 January 2026
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 10  (11) , 2404-2415
Rotational ultrasound and photoacoustic tomography of the human body
Imaging the human body′s morphological and angiographic information is essential for diagnosing, monitoring and treating medical conditions. Here we combine the power of ultrasonography for morphological assessment of soft tissue with photoacoustic tomography (PAT) for visualizing blood vessels to enable three-dimensional (3D) panoramic imaging. Specifically, fast panoramic rotational ultrasound tomography and PAT are integrated for hybrid rotational ultrasound and photoacoustic tomography (RUS-PAT), which obtains 3D ultrasound structural and PAT angiographic images of the human body quasi-simultaneously. The rotational ultrasound tomography functionality is achieved using a single-element ultrasonic transducer for ultrasound transmission and rotating arc-shaped arrays for 3D panoramic detection. By switching the acoustic source to a light source, the system is conveniently converted to PAT mode to acquire angiographic images in the same region. Using RUS-PAT, we successfully imaged the human head, breast, hand and foot with a 10-cm-diameter field of view, submillimetre isotropic resolution and 10 s imaging time for each modality. 3D RUS-PAT is a powerful tool for high-speed, dual-contrast imaging of the human body with potential for rapid clinical translation.
- 16 January 2026
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 10  (11) , 2404-2415
Top Cited PapersRotational ultrasound and photoacoustic tomography of the human body
Imaging the human body′s morphological and angiographic information is essential for diagnosing, monitoring and treating medical conditions. Here we combine the power of ultrasonography for morphological assessment of soft tissue with photoacoustic tomography (PAT) for visualizing blood vessels to enable three-dimensional (3D) panoramic imaging. Specifically, fast panoramic rotational ultrasound tomography and PAT are integrated for hybrid rotational ultrasound and photoacoustic tomography (RUS-PAT), which obtains 3D ultrasound structural and PAT angiographic images of the human body quasi-simultaneously. The rotational ultrasound tomography functionality is achieved using a single-element ultrasonic transducer for ultrasound transmission and rotating arc-shaped arrays for 3D panoramic detection. By switching the acoustic source to a light source, the system is conveniently converted to PAT mode to acquire angiographic images in the same region. Using RUS-PAT, we successfully imaged the human head, breast, hand and foot with a 10-cm-diameter field of view, submillimetre isotropic resolution and 10 s imaging time for each modality. 3D RUS-PAT is a powerful tool for high-speed, dual-contrast imaging of the human body with potential for rapid clinical translation.
- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 45  (5) , 2015-2022
Multifocal Optical-Resolution Photoacoustic Microscopy With a Masked Single-Element Transducer
Optical-resolution photoacoustic microscopy (OR-PAM) can visualize cellular-level wavelength-dependent optical absorption with high resolution and sensitivity. However, the imaging speed of OR-PAM has been limited by the laser repetition rate due to the point-by-point scanning of a focused laser beam. To overcome this limitation, we propose multifocal optical-resolution photoacoustic microscopy (MOR-PAM) with a single-element ultrasonic transducer, leveraging the diffractive optical element (DOE) and a custom-designed encoding acoustic mask. The DOE generates 8 focal spots of 3 ÎĽm diameter. The acoustic mask was designed to encode photoacoustic signals from different focal spots. MOR-PAM achieved an 8-fold increase in imaging speed compared to conventional OR-PAM with the same laser repetition rate. We demonstrated the MOR-PAM using a 266 nm laser at 10 KHz, providing solutions for rapid OR-PAM beyond the laser repetition rate in a cost-effective way. The proposed method can be applied to versatile OR-PAM configurations and enable new applications where high-speed imaging is critical.
- 12 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 45  (5) , 2015-2022
Multifocal Optical-Resolution Photoacoustic Microscopy With a Masked Single-Element Transducer
Optical-resolution photoacoustic microscopy (OR-PAM) can visualize cellular-level wavelength-dependent optical absorption with high resolution and sensitivity. However, the imaging speed of OR-PAM has been limited by the laser repetition rate due to the point-by-point scanning of a focused laser beam. To overcome this limitation, we propose multifocal optical-resolution photoacoustic microscopy (MOR-PAM) with a single-element ultrasonic transducer, leveraging the diffractive optical element (DOE) and a custom-designed encoding acoustic mask. The DOE generates 8 focal spots of 3 ÎĽm diameter. The acoustic mask was designed to encode photoacoustic signals from different focal spots. MOR-PAM achieved an 8-fold increase in imaging speed compared to conventional OR-PAM with the same laser repetition rate. We demonstrated the MOR-PAM using a 266 nm laser at 10 KHz, providing solutions for rapid OR-PAM beyond the laser repetition rate in a cost-effective way. The proposed method can be applied to versatile OR-PAM configurations and enable new applications where high-speed imaging is critical.
- 22 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 45  (3) , 906-918
Data-Driven System Matrix Manipulation Enabling Fast Functional Imaging in Tomography
Tomographic imaging modalities are described by large system matrices. To improve the temporal resolution of functional imaging in tomography, sparse spatial sampling is often employed, which degrades the system matrix and introduces artifacts in reconstructed images. Various existing techniques improve the image quality without correcting the system matrix and have limitations. Here, we compress the system matrix to improve computational efficiency (e.g., 42 times) using singular value decomposition and fast Fourier transform. Enabled by the efficiency, we propose fast sparsely sampling functional imaging by incorporating a densely sampled prior image into the system matrix, which maintains the critical linearity while mitigating artifacts. We demonstrate the methods in 3D photoacoustic computed tomography with significantly improved image quality and clarify their applicability to X-ray CT and radial-sampling MRI due to the similarities in system matrices.
- 22 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 45  (3) , 906-918
Data-Driven System Matrix Manipulation Enabling Fast Functional Imaging in Tomography
Tomographic imaging modalities are described by large system matrices. To improve the temporal resolution of functional imaging in tomography, sparse spatial sampling is often employed, which degrades the system matrix and introduces artifacts in reconstructed images. Various existing techniques improve the image quality without correcting the system matrix and have limitations. Here, we compress the system matrix to improve computational efficiency (e.g., 42 times) using singular value decomposition and fast Fourier transform. Enabled by the efficiency, we propose fast sparsely sampling functional imaging by incorporating a densely sampled prior image into the system matrix, which maintains the critical linearity while mitigating artifacts. We demonstrate the methods in 3D photoacoustic computed tomography with significantly improved image quality and clarify their applicability to X-ray CT and radial-sampling MRI due to the similarities in system matrices.
- 9 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44  (2) , 645-655
Full-Wave Image Reconstruction in Transcranial Photoacoustic Computed Tomography Using a Finite Element Method
Transcranial photoacoustic computed tomography presents challenges in human brain imaging due to skull-induced acoustic aberration. Existing full-wave image reconstruction methods rely on a unified elastic wave equation for skull shear and longitudinal wave propagation, therefore demanding substantial computational resources. We propose an efficient discrete imaging model based on finite element discretization. The elastic wave equation for solids is solely applied to the hard-tissue skull region, while the soft-tissue or coupling-medium region that dominates the simulation domain is modeled with the simpler acoustic wave equation for liquids. The solid-liquid interfaces are explicitly modeled with elastic-acoustic coupling. Furthermore, finite element discretization allows coarser, irregular meshes to conform to object geometry. These factors significantly reduce the linear system size by 20 times to facilitate accurate whole-brain simulations with improved speed. We derive a matched forward-adjoint operator pair based on the model to enable integration with various optimization algorithms. We validate the reconstruction framework through numerical simulations and phantom experiments.
- 9 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44  (2) , 645-655
Full-Wave Image Reconstruction in Transcranial Photoacoustic Computed Tomography Using a Finite Element Method
Transcranial photoacoustic computed tomography presents challenges in human brain imaging due to skull-induced acoustic aberration. Existing full-wave image reconstruction methods rely on a unified elastic wave equation for skull shear and longitudinal wave propagation, therefore demanding substantial computational resources. We propose an efficient discrete imaging model based on finite element discretization. The elastic wave equation for solids is solely applied to the hard-tissue skull region, while the soft-tissue or coupling-medium region that dominates the simulation domain is modeled with the simpler acoustic wave equation for liquids. The solid-liquid interfaces are explicitly modeled with elastic-acoustic coupling. Furthermore, finite element discretization allows coarser, irregular meshes to conform to object geometry. These factors significantly reduce the linear system size by 20 times to facilitate accurate whole-brain simulations with improved speed. We derive a matched forward-adjoint operator pair based on the model to enable integration with various optimization algorithms. We validate the reconstruction framework through numerical simulations and phantom experiments.
- 21 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 71  (10) , 2948-2955
Quantification of Cervical Elasticity During Pregnancy Based on Transvaginal Ultrasound Imaging and Stress Measurement
Strain elastography and shear wave elastography are commonly used to quantify cervical elasticity. However, the absence of stress information in strain elastography causes difficulty in inter-session elasticity comparison, and the robustness of shear wave elastography is compromised by cervical tissue's high inhomogeneity. Objective: To overcome these limitations, we develop a quantitative cervical elastography system by adding a stress sensor to a clinically used transvaginal ultrasound imaging system. Methods: We record the cervical deformation in B-mode images and measure the probe-surface stress through the sensor. Then we quantify the strain using a customized algorithm and estimate the cervical Young's modulus through stress-strain linear regression. Results: In phantom experiments, we demonstrate the system's high accuracy (alignment with the quasi-static compression method, p -value = 0.369 > 0.05), robustness (alignment between 60°- and 90°-contact measurements, p -value = 0.638 > 0.05), repeatability (consistency of single sonographers' measurements, coefficient of variation < 0.06), and reproducibility (alignment between two sonographers' measurements, Pearson correlation coefficient = 0.981). Applying it to pregnant participants, we observe significant cervical softening ( p -value < 0.001): Young's modulus decreases 3.95% weekly and its geometric mean value during the first (11 to 13 weeks), second, and third trimesters are 13.07 kPa, 7.59 kPa, and 4.40 kPa, respectively. Conclusion: The proposed system is accurate, robust, and safe, and enables longitudinal and inter-examiner comparisons. Significance: The system applies to different ultrasound machines with minor software updates, which allows for studies of cervical softening patterns in pregnancy for larger populations, facilitating insights into conditions such as preterm birth.
- 21 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 71  (10) , 2948-2955
Quantification of Cervical Elasticity During Pregnancy Based on Transvaginal Ultrasound Imaging and Stress Measurement
Strain elastography and shear wave elastography are commonly used to quantify cervical elasticity. However, the absence of stress information in strain elastography causes difficulty in inter-session elasticity comparison, and the robustness of shear wave elastography is compromised by cervical tissue's high inhomogeneity. Objective: To overcome these limitations, we develop a quantitative cervical elastography system by adding a stress sensor to a clinically used transvaginal ultrasound imaging system. Methods: We record the cervical deformation in B-mode images and measure the probe-surface stress through the sensor. Then we quantify the strain using a customized algorithm and estimate the cervical Young's modulus through stress-strain linear regression. Results: In phantom experiments, we demonstrate the system's high accuracy (alignment with the quasi-static compression method, p -value = 0.369 > 0.05), robustness (alignment between 60°- and 90°-contact measurements, p -value = 0.638 > 0.05), repeatability (consistency of single sonographers' measurements, coefficient of variation < 0.06), and reproducibility (alignment between two sonographers' measurements, Pearson correlation coefficient = 0.981). Applying it to pregnant participants, we observe significant cervical softening ( p -value < 0.001): Young's modulus decreases 3.95% weekly and its geometric mean value during the first (11 to 13 weeks), second, and third trimesters are 13.07 kPa, 7.59 kPa, and 4.40 kPa, respectively. Conclusion: The proposed system is accurate, robust, and safe, and enables longitudinal and inter-examiner comparisons. Significance: The system applies to different ultrasound machines with minor software updates, which allows for studies of cervical softening patterns in pregnancy for larger populations, facilitating insights into conditions such as preterm birth.
- Open Access
Data-driven system matrix manipulation enabling fast functional imaging and intra-image nonrigid motion correction in tomography
Tomographic imaging modalities are described by large system matrices. Sparse sampling and tissue motion degrade system matrix and image quality. Various existing techniques improve the image quality without correcting the system matrices. Here, we compress the system matrices to improve computational efficiency (e.g., 42 times) using singular value decomposition and fast Fourier transform. Enabled by the efficiency, we propose (1) fast sparsely sampling functional imaging by incorporating a densely sampled prior image into the system matrix, which maintains the critical linearity while mitigating artifacts and (2) intra-image nonrigid motion correction by incorporating the motion as subdomain translations into the system matrix and reconstructing the translations together with the image iteratively. We demonstrate the methods in 3D photoacoustic computed tomography with significantly improved image qualities and clarify their applicability to X-ray CT and MRI or other types of imperfections due to the similarities in system matrices.
Open AccessData-driven system matrix manipulation enabling fast functional imaging and intra-image nonrigid motion correction in tomography
Tomographic imaging modalities are described by large system matrices. Sparse sampling and tissue motion degrade system matrix and image quality. Various existing techniques improve the image quality without correcting the system matrices. Here, we compress the system matrices to improve computational efficiency (e.g., 42 times) using singular value decomposition and fast Fourier transform. Enabled by the efficiency, we propose (1) fast sparsely sampling functional imaging by incorporating a densely sampled prior image into the system matrix, which maintains the critical linearity while mitigating artifacts and (2) intra-image nonrigid motion correction by incorporating the motion as subdomain translations into the system matrix and reconstructing the translations together with the image iteratively. We demonstrate the methods in 3D photoacoustic computed tomography with significantly improved image qualities and clarify their applicability to X-ray CT and MRI or other types of imperfections due to the similarities in system matrices.
- 23 December 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83  (40) , 88303-88322
Thermal infrared and visible sequences tracking via dual adversarial pixel fusion
- 23 December 2023
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 83  (40) , 88303-88322
Thermal infrared and visible sequences tracking via dual adversarial pixel fusion
- 30 November 2023
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 8  (6) , 712-725
Ultrafast longitudinal imaging of haemodynamics via single-shot volumetric photoacoustic tomography with a single-element detector
Techniques for imaging haemodynamics use ionizing radiation or contrast agents or are limited by imaging depth (within approximately 1 mm), complex and expensive data-acquisition systems, or low imaging speeds, system complexity or cost. Here we show that ultrafast volumetric photoacoustic imaging of haemodynamics in the human body at up to 1 kHz can be achieved using a single laser pulse and a single element functioning as 6,400 virtual detectors. The technique, which does not require recalibration for different objects or during long-term operation, enables the longitudinal volumetric imaging of haemodynamics in vasculature a few millimetres below the skin’s surface. We demonstrate this technique in vessels in the feet of healthy human volunteers by capturing haemodynamic changes in response to vascular occlusion. Single-shot volumetric photoacoustic imaging using a single-element detector may facilitate the early detection and monitoring of peripheral vascular diseases and may be advantageous for use in biometrics and point-of-care testing.
- 30 November 2023
- Published by Springer Nature in Nature Biomedical Engineering
- Vol. 8  (6) , 712-725
Ultrafast longitudinal imaging of haemodynamics via single-shot volumetric photoacoustic tomography with a single-element detector
Techniques for imaging haemodynamics use ionizing radiation or contrast agents or are limited by imaging depth (within approximately 1 mm), complex and expensive data-acquisition systems, or low imaging speeds, system complexity or cost. Here we show that ultrafast volumetric photoacoustic imaging of haemodynamics in the human body at up to 1 kHz can be achieved using a single laser pulse and a single element functioning as 6,400 virtual detectors. The technique, which does not require recalibration for different objects or during long-term operation, enables the longitudinal volumetric imaging of haemodynamics in vasculature a few millimetres below the skin’s surface. We demonstrate this technique in vessels in the feet of healthy human volunteers by capturing haemodynamic changes in response to vascular occlusion. Single-shot volumetric photoacoustic imaging using a single-element detector may facilitate the early detection and monitoring of peripheral vascular diseases and may be advantageous for use in biometrics and point-of-care testing.
- 20 July 2023
- Published by Elsevier in Journal of Hepatology
- Vol. 78, S1112
HBV infected pregnant woman received antiviral treatment and discontinued it postpartum can benefit from retreatment
- 20 July 2023
- Published by Elsevier in Journal of Hepatology
- Vol. 78, S1112
HBV infected pregnant woman received antiviral treatment and discontinued it postpartum can benefit from retreatment
- Open Access
- 14 July 2023
- Published by Springer Nature in Research Square
Visible and Thermal Infrared Light Target Tracking via Multi-Adapter Network with Feature-Wise-Transformations
Existing deep learning-based RGBT target tracking algorithms focus on extracting the features of each modal image separately by using the same CNN model, and then stitching the extracted features to achieve the fusion of multimodal data. However, the reliability of single-modal visible light images and infrared light images will degrades with time shift, and when one of the modal data is unreliable, the related operations of feature extraction and fusion of the two modal data tend to introduce discriminative information that impairs tracking robustness, and makes it easy to lose the initial target in long-time tracking. To this end, we design a Feature-wise transformations network for feature modulation for RGBT target tracking to ensure that the information fusion between the two modalities is as complementary as possible to the strengths of the multimodal data. Most of the existing RGBT tracking algorithms focus on the fusion of multimodal information, and the selection of the target frame mainly depends on the classification confidence of classifying the object as the target or background. Due to the classification confidence and the localization confidence may not match, it is easy to cause inaccurate target localization. In order to improve the localization accuracy of the tracking process, we optimize the instance adapter in the MANet al.gorithm with a new method of updating the classification confidence based on the sample box with the highest localization confidence, and design a new method of updating the localization confidence based on the sample box with the highest classification confidence to obtain the final candidate sample box for bounding box regression. The three improvements are applied to the MANet al.gorithm to construct a new RGBT target tracking network, and the test results on three benchmark datasets of RGBT target tracking show that our improvements are effective in improving the accuracy and success rate of the tracker.
- 14 July 2023
- Published by Springer Nature in Research Square
Open AccessVisible and Thermal Infrared Light Target Tracking via Multi-Adapter Network with Feature-Wise-Transformations
Existing deep learning-based RGBT target tracking algorithms focus on extracting the features of each modal image separately by using the same CNN model, and then stitching the extracted features to achieve the fusion of multimodal data. However, the reliability of single-modal visible light images and infrared light images will degrades with time shift, and when one of the modal data is unreliable, the related operations of feature extraction and fusion of the two modal data tend to introduce discriminative information that impairs tracking robustness, and makes it easy to lose the initial target in long-time tracking. To this end, we design a Feature-wise transformations network for feature modulation for RGBT target tracking to ensure that the information fusion between the two modalities is as complementary as possible to the strengths of the multimodal data. Most of the existing RGBT tracking algorithms focus on the fusion of multimodal information, and the selection of the target frame mainly depends on the classification confidence of classifying the object as the target or background. Due to the classification confidence and the localization confidence may not match, it is easy to cause inaccurate target localization. In order to improve the localization accuracy of the tracking process, we optimize the instance adapter in the MANet al.gorithm with a new method of updating the classification confidence based on the sample box with the highest localization confidence, and design a new method of updating the localization confidence based on the sample box with the highest classification confidence to obtain the final candidate sample box for bounding box regression. The three improvements are applied to the MANet al.gorithm to construct a new RGBT target tracking network, and the test results on three benchmark datasets of RGBT target tracking show that our improvements are effective in improving the accuracy and success rate of the tracker.