Jyh-Cheng Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Jyh-Cheng Chen has a total of 306 co-authors. Here are the top co-authors by number of shared publications:
- Shih-Chun Jin
National Yang-Ming University
9 shared publications Ren-Shyan Liu
National Yang-Ming University, Department of Biomedical Imaging and Radiological Sciences, Taipei 112, Taiwan
8 shared publications- Yi Ren
University of East Anglia
6 shared publications Shyh-Jen Wang
Department of Nuclear Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
5 shared publications- Chien Chen
National Yang Ming Chiao Tung University, Hsinchu, Taiwan
4 shared publications Ho-Shiang Chueh
4 shared publications
Latest Publications
- 21 July 2026
- Published by Wiley in Laser & Photonics Reviews
- p. e71613
Dual Response of a Perovskite Scintillator for Low‑Dose X‑Ray Imaging Detection and Beam Monitoring in Proton Radiotherapy
The development of novel high‐performance radiation detectors is crucial for medical applications, particularly in X‐ray diagnostics and proton radiotherapy. Herein, we report a low‐cost perovskite single‐crystal scintillator as a versatile and scalable candidate for the detection of both x‐rays and high‐energy proton beams. Theas‐prepared Cs 3 Cu 2 I 5 : 20%Mn scintillator exhibits outstanding response characteristics across the energy ranges relevant to X‐ray medical imaging (50–150 keV) and proton therapy (70–220 MeV). It achieves a high light yield of 55 260 photons/MeV, and its response signals show excellent linearity with respect to both x‐ray dose and proton flux. Moreover, it exhibits a sensitive dynamic response to proton beams with different energies, achieving a minimum detection limit of 1.34 × 10 4 p/s. Reliability and environmental stability tests further confirm its potential for long‐term application in proton beam detection. Furthermore, a flexible film based on this scintillator enables real‐time, high‐definition visualization of proton beam spots and realizes dynamic beam tracking, verifying its capability for real‐time proton beam monitoring. This work highlights the great potential of Mn 2+ ‐doped Cs 3 Cu 2 I 5 as a favorable‐performance, wide‐range, and low‐cost material for medical radiation detection and provides new experimental insights for advancing perovskite‐based detectors in clinical proton therapy and radiography.
- 21 July 2026
- Published by Wiley in Laser & Photonics Reviews
- p. e71613
Dual Response of a Perovskite Scintillator for Low‑Dose X‑Ray Imaging Detection and Beam Monitoring in Proton Radiotherapy
The development of novel high‐performance radiation detectors is crucial for medical applications, particularly in X‐ray diagnostics and proton radiotherapy. Herein, we report a low‐cost perovskite single‐crystal scintillator as a versatile and scalable candidate for the detection of both x‐rays and high‐energy proton beams. Theas‐prepared Cs 3 Cu 2 I 5 : 20%Mn scintillator exhibits outstanding response characteristics across the energy ranges relevant to X‐ray medical imaging (50–150 keV) and proton therapy (70–220 MeV). It achieves a high light yield of 55 260 photons/MeV, and its response signals show excellent linearity with respect to both x‐ray dose and proton flux. Moreover, it exhibits a sensitive dynamic response to proton beams with different energies, achieving a minimum detection limit of 1.34 × 10 4 p/s. Reliability and environmental stability tests further confirm its potential for long‐term application in proton beam detection. Furthermore, a flexible film based on this scintillator enables real‐time, high‐definition visualization of proton beam spots and realizes dynamic beam tracking, verifying its capability for real‐time proton beam monitoring. This work highlights the great potential of Mn 2+ ‐doped Cs 3 Cu 2 I 5 as a favorable‐performance, wide‐range, and low‐cost material for medical radiation detection and provides new experimental insights for advancing perovskite‐based detectors in clinical proton therapy and radiography.
- Open Access
- 18 May 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (3)
Polychromatic sinogram based material decomposition and reconstruction methods: Integrate standard reference database with an in-house cone-beam micro-computed tomography system
This study develops a sinogram-based material decomposition framework for a dual-energy cone-beam micro-CT system, providing quantitative spectral imaging without specialized photon-counting detectors. The aim is to address challenges in material characterization, particularly for preclinical in-vivo imaging. The methodology involves acquiring dual-energy projections (50 and 80 kVp) of a standardized reference phantom to calibrate a three-material known basis model. This model is then solved in the sinogram domain using a robust and computationally efficient Moore-Penrose pseudoinverse solver to generate basis material maps. By integrating these maps with a reference database, the framework synthesizes artifact-reduced virtual monochromatic images by photoelectric effect images, and Compton scatter images. Our results demonstrate that the framework, calibrated once, is highly generalizable and can be reliably applied to other phantoms and complex in-vivo rat scans. The method achieved accurate, automated segmentation of skeletal structures from soft tissues in in-vivo data, which was quantitatively validated on phantoms with structural similarity index values approaching 1.0. The synthesis successfully separated physical attenuation mechanisms, isolating bone-dominant signals in PEIs and soft-tissue signals in CSIs. In conclusion, this study validates an accessible and robust sinogram-based MD method that enables quantitative, multi-material analysis on conventional micro-CT systems, offering a practical tool for advanced preclinical research.
- 18 May 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (3)
Open AccessPolychromatic sinogram based material decomposition and reconstruction methods: Integrate standard reference database with an in-house cone-beam micro-computed tomography system
This study develops a sinogram-based material decomposition framework for a dual-energy cone-beam micro-CT system, providing quantitative spectral imaging without specialized photon-counting detectors. The aim is to address challenges in material characterization, particularly for preclinical in-vivo imaging. The methodology involves acquiring dual-energy projections (50 and 80 kVp) of a standardized reference phantom to calibrate a three-material known basis model. This model is then solved in the sinogram domain using a robust and computationally efficient Moore-Penrose pseudoinverse solver to generate basis material maps. By integrating these maps with a reference database, the framework synthesizes artifact-reduced virtual monochromatic images by photoelectric effect images, and Compton scatter images. Our results demonstrate that the framework, calibrated once, is highly generalizable and can be reliably applied to other phantoms and complex in-vivo rat scans. The method achieved accurate, automated segmentation of skeletal structures from soft tissues in in-vivo data, which was quantitatively validated on phantoms with structural similarity index values approaching 1.0. The synthesis successfully separated physical attenuation mechanisms, isolating bone-dominant signals in PEIs and soft-tissue signals in CSIs. In conclusion, this study validates an accessible and robust sinogram-based MD method that enables quantitative, multi-material analysis on conventional micro-CT systems, offering a practical tool for advanced preclinical research.
- Open Access
- 23 April 2026
- Published by MDPI AG in Diagnostics
- Vol. 16 (9) , 1270
Epicardial Adipose Tissue CT Radiomics Improves Acute Coronary Syndrome Prediction Beyond Coronary Artery Calcium Score
Objectives: To determine if global epicardial adipose tissue (EAT) radiomics, derived from non-contrast coronary artery calcium (CAC) scans, improves acute coronary syndrome (ACS) prediction beyond traditional risk factors (TRFs) and Agatston score (AS) in individuals without angina. Methods: We retrospectively analyzed 2020 subjects without angina who underwent CAC scans from 2016 to 2019, among whom 76 patients developed acute coronary syndrome (ACS) during a follow-up period until December 2023. One-to-one propensity score matching (PSM) based on TRFs (age, sex, BMI, smoking, diabetes, hypertension, dyslipidemia) and Agatston score ranks (ASR) created 76 ACS and 76 matched non-ACS subjects. A radiomics model was built using 5-fold cross-validation on the matched cohort and tested on the entire unmatched cohort. Statistical tests included AUC comparison. Results: PSM effectively mitigated disparities of TRFs and ASR. The radiomics model achieved AUCs of 0.91 ± 0.01 and 0.89 ± 0.03 for the training and matched test sets, respectively, and 0.89 ± 0.03 on the unmatched cohort. The radiomics score significantly improved ACS prediction over TRF and CAC models (p < 0.001) in both matched and unmatched cohorts. Conclusions: Global EAT radiomic phenotypes from conventional CAC scan improve ACS risk stratification beyond TRFs and AS in individuals without angina. In contrast to PCAT on CCTA, our simple approach appears to be suitable for large-scale applications using preventative CAC scans.
- 23 April 2026
- Published by MDPI AG in Diagnostics
- Vol. 16 (9) , 1270
Open AccessEpicardial Adipose Tissue CT Radiomics Improves Acute Coronary Syndrome Prediction Beyond Coronary Artery Calcium Score
Objectives: To determine if global epicardial adipose tissue (EAT) radiomics, derived from non-contrast coronary artery calcium (CAC) scans, improves acute coronary syndrome (ACS) prediction beyond traditional risk factors (TRFs) and Agatston score (AS) in individuals without angina. Methods: We retrospectively analyzed 2020 subjects without angina who underwent CAC scans from 2016 to 2019, among whom 76 patients developed acute coronary syndrome (ACS) during a follow-up period until December 2023. One-to-one propensity score matching (PSM) based on TRFs (age, sex, BMI, smoking, diabetes, hypertension, dyslipidemia) and Agatston score ranks (ASR) created 76 ACS and 76 matched non-ACS subjects. A radiomics model was built using 5-fold cross-validation on the matched cohort and tested on the entire unmatched cohort. Statistical tests included AUC comparison. Results: PSM effectively mitigated disparities of TRFs and ASR. The radiomics model achieved AUCs of 0.91 ± 0.01 and 0.89 ± 0.03 for the training and matched test sets, respectively, and 0.89 ± 0.03 on the unmatched cohort. The radiomics score significantly improved ACS prediction over TRF and CAC models (p < 0.001) in both matched and unmatched cohorts. Conclusions: Global EAT radiomic phenotypes from conventional CAC scan improve ACS risk stratification beyond TRFs and AS in individuals without angina. In contrast to PCAT on CCTA, our simple approach appears to be suitable for large-scale applications using preventative CAC scans.
- 1 April 2026
- Published by Springer Nature
- p. 237-245
A Siamese Vision Transfer Architecture for Prediction of Brain Tumour Responses during Therapy
- 1 April 2026
- Published by Springer Nature
- p. 237-245
A Siamese Vision Transfer Architecture for Prediction of Brain Tumour Responses during Therapy
- 29 January 2026
- Published by Wolters Kluwer Health in Nuclear Medicine Communications
- Vol. 47 (5) , 552-562
Refining deep learning segmentation in gallium-68-prostate-specific membrane antigen-11 positron emission tomography: evaluation of small lesion filtering and intersection-over-union thresholds
Purpose: Evaluate how small lesion filtering and different intersection-over-union (IoU) thresholds influence deep learning segmentation of Ga-68-PSMA-11 PET images in prostate cancer. Methods: A 3D U-Net was trained on 115 patient scans with manual contours as ground truth. Performance was assessed at voxel, lesion and patient levels. Lesions less than 8 voxels (195 mm 3 ) or less than 27 voxels (658 mm 3 ) were optionally discarded, and lesion-level metrics were computed across 10–50% IoU thresholds. Results: Excluding lesions less than 27 voxels increased voxel-level Dice from 0.7975 to 0.8173 and improved precision. Lesion-level metrics were stable across IoU thresholds of 20–40% after 27-voxel filtering, while sensitivity declined at higher overlap thresholds. Patient-level sensitivity and positive predictive value reached 96.6 and 94.5%, respectively. Conclusion: Implementing small lesion filtering criteria improves the reliability of automated segmentation outputs, particularly for quantitative evaluation. For lesion-level metrics, defining true positives within the range of 20–40% IoU thresholds is optimal. Further validation through multicenter studies and larger datasets is essential to ensure the generalizability of these findings.
- 29 January 2026
- Published by Wolters Kluwer Health in Nuclear Medicine Communications
- Vol. 47 (5) , 552-562
Refining deep learning segmentation in gallium-68-prostate-specific membrane antigen-11 positron emission tomography: evaluation of small lesion filtering and intersection-over-union thresholds
Purpose: Evaluate how small lesion filtering and different intersection-over-union (IoU) thresholds influence deep learning segmentation of Ga-68-PSMA-11 PET images in prostate cancer. Methods: A 3D U-Net was trained on 115 patient scans with manual contours as ground truth. Performance was assessed at voxel, lesion and patient levels. Lesions less than 8 voxels (195 mm 3 ) or less than 27 voxels (658 mm 3 ) were optionally discarded, and lesion-level metrics were computed across 10–50% IoU thresholds. Results: Excluding lesions less than 27 voxels increased voxel-level Dice from 0.7975 to 0.8173 and improved precision. Lesion-level metrics were stable across IoU thresholds of 20–40% after 27-voxel filtering, while sensitivity declined at higher overlap thresholds. Patient-level sensitivity and positive predictive value reached 96.6 and 94.5%, respectively. Conclusion: Implementing small lesion filtering criteria improves the reliability of automated segmentation outputs, particularly for quantitative evaluation. For lesion-level metrics, defining true positives within the range of 20–40% IoU thresholds is optimal. Further validation through multicenter studies and larger datasets is essential to ensure the generalizability of these findings.
- 9 December 2025
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (4) , 3573-3591
A Pilot Study: PET/CT Cross-Modal-Based Multi-Head Fusion Attention Generative Adversarial Network (MHFA-GAN) for PET Image Super-Resolution
Positron emission tomography (PET) is a valuable imaging technique for assessing cellular metabolic activity and is widely used for early disease detection and diagnosis. While high-resolution (HR) PET images offer richer diagnostic information, achieving such quality often requires expensive equipment and time-intensive reconstruction processes. To address these challenges without hardware upgrades, this paper introduces a novel multi-head fusion attention generative adversarial network (MHFA-GAN), designed to enhance the spatial resolution and reconstruction efficiency of PET images through super-resolution (SR) techniques. The MHFA-GAN employs a multi-head mixed convolution mechanism to extract complementary features from PET and computed tomography (CT) images. An enhanced multi-head fusion attention module (EMHFA) is incorporated to adaptively adjust feature weights and align cross-modal information, effectively minimizing the impact of sub-pixel displacement while integrating metabolic and anatomical details. During the reconstruction phase, a global residual block (GRB) is utilized to balance local high-frequency details with global contextual information. Experiments conducted on small animal (SA), phantom, capillary, and 22Na point source datasets demonstrate that MHFA-GAN outperforms current state-of-the-art methods in both qualitative and quantitative evaluations. Additionally, the model also shows significant results in SR tasks on different device datasets, further verifying its cross-device applicability and robustness. Using the full width at half maximum (FWHM) metric to assess spatial resolution, the SR images from the capillary dataset achieved a FWHM value of 0.625 mm, approaching the physical limit of the SA PET scanner (0.6 mm) and significantly outperforming the HR images (0.871 mm). These results highlight the effectiveness of MHFA-GAN in improving PET image quality, offering a cost-effective solution to enhance spatial resolution and support clinical diagnostics.
- 9 December 2025
- Published by Springer Nature in Journal of Imaging Informatics in Medicine
- Vol. 39 (4) , 3573-3591
A Pilot Study: PET/CT Cross-Modal-Based Multi-Head Fusion Attention Generative Adversarial Network (MHFA-GAN) for PET Image Super-Resolution
Positron emission tomography (PET) is a valuable imaging technique for assessing cellular metabolic activity and is widely used for early disease detection and diagnosis. While high-resolution (HR) PET images offer richer diagnostic information, achieving such quality often requires expensive equipment and time-intensive reconstruction processes. To address these challenges without hardware upgrades, this paper introduces a novel multi-head fusion attention generative adversarial network (MHFA-GAN), designed to enhance the spatial resolution and reconstruction efficiency of PET images through super-resolution (SR) techniques. The MHFA-GAN employs a multi-head mixed convolution mechanism to extract complementary features from PET and computed tomography (CT) images. An enhanced multi-head fusion attention module (EMHFA) is incorporated to adaptively adjust feature weights and align cross-modal information, effectively minimizing the impact of sub-pixel displacement while integrating metabolic and anatomical details. During the reconstruction phase, a global residual block (GRB) is utilized to balance local high-frequency details with global contextual information. Experiments conducted on small animal (SA), phantom, capillary, and 22Na point source datasets demonstrate that MHFA-GAN outperforms current state-of-the-art methods in both qualitative and quantitative evaluations. Additionally, the model also shows significant results in SR tasks on different device datasets, further verifying its cross-device applicability and robustness. Using the full width at half maximum (FWHM) metric to assess spatial resolution, the SR images from the capillary dataset achieved a FWHM value of 0.625 mm, approaching the physical limit of the SA PET scanner (0.6 mm) and significantly outperforming the HR images (0.871 mm). These results highlight the effectiveness of MHFA-GAN in improving PET image quality, offering a cost-effective solution to enhance spatial resolution and support clinical diagnostics.
- Open Access
- 19 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 4928-4929
free5GC '25: The 1st free5GC World Forum
- 19 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 4928-4929
Open Accessfree5GC '25: The 1st free5GC World Forum
- Open Access
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-9
A Machine Learning-Driven NWDAF Architecture For Intelligent 5G Core Networks
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-9
Open AccessA Machine Learning-Driven NWDAF Architecture For Intelligent 5G Core Networks
- Open Access
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Creation and Support of 5G Core Network Slices with a focus on eMBB
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessCreation and Support of 5G Core Network Slices with a focus on eMBB
- Open Access
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Rethinking 5G Core for Performance, Fairness, and Flexibility
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Open AccessRethinking 5G Core for Performance, Fairness, and Flexibility