Physics in medicine and biology最新文献

{"title":"Simulating atherosclerotic plaque mechanics using polyvinyl alcohol (PVA) cryogel artery phantoms, ultrasound imaging and inverse finite element analysis.","authors":"Yasmine Guendouz, Noor Adeebah Mohamed Razif, Floriane Bernasconi, Gordon O' Brien, Robert D Johnston, Caitríona Lally","doi":"10.1088/1361-6560/ad9a4b","DOIUrl":"10.1088/1361-6560/ad9a4b","url":null,"abstract":"<p><p>The clinical decision to establish if a patient with carotid disease should undergo surgical intervention is primarily based on the percent stenosis. Whilst this applies for high-grade stenosed vessels (>70%), it falls short for other cases. Due to the heterogeneity of plaque tissue, probing the mechanics of the tissue would likely provide further insights into why some plaques are more prone to rupture. Mechanical characterization of such tissue is nontrivial, however, due to the difficulties in collecting fresh, intact plaque tissue and using physiologically relevant mechanical testing of such material. The use of polyvinyl alcohol (PVA) cryogel is thus highly convenient because of its acoustic properties and tunable mechanical properties.<i>Methods.</i>The aim of this study is to demonstrate the potential of PVA phantoms to simulate atherosclerotic features. In addition, a testing and simulation framework is developed for full PVA vessel material characterization using ring tensile testing and inflation testing combined with non-invasive ultrasound imaging and computational modeling.<i>Results.</i>Strain stiffening behavior was observed in PVA through ring tensile tests, particularly at high (<i>n</i>= 6) freeze-thaw cycles (FTCs). Inflation testing of bi-layered phantoms featuring lipid pool inclusions demonstrated high strains at shoulder regions. The application of an inverse finite element framework successfully recovered boundaries and determined the shear moduli for the PVA wall to lie within the range 27-53 kPa.<i>Conclusion.</i>The imaging-modeling framework presented facilitates the use and characterization of arterial mimicking phantoms to further explore plaque rupture. It also shows translational potential for non-invasive mechanical characterization of atherosclerotic plaques to improve the identification of clinically relevant metrics of plaque vulnerability.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142771349","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring the performance of a DOI-capable TOF-PET module using different SiPMs, customized and commercial readout electronics.","authors":"Giulia Terragni, Vanessa Nadig, Elena Tribbia, Stefano Di Gangi, Ekaterini Toumparidou, Thomas Meyer, Johann Marton, Volkmar Schulz, Stefan Gundacker, Marco Pizzichemi, Etiennette Auffray","doi":"10.1088/1361-6560/ada19a","DOIUrl":"https://doi.org/10.1088/1361-6560/ada19a","url":null,"abstract":"<p><strong>Objective: </strong>Time resolution is crucial in positron emission tomography (PET) to enhance the signal-to-noise ratio and image quality. Moreover, high sensitivity requires long scintillators, which can cause distortions in the reconstructed images due to parallax effects. This study evaluates the performance of a time-of-flight (TOF)-PET module that makes use of a single-side readout of a 4x4 3.1x3.1x15 mm³LYSO:Ce matrix with an array of 4x4 SiPMs and a light guide to extract high-resolution TOF and depth of interaction (DOI) information. &#xD;Approach: This study assesses the performance of the detector prototype using the commercially available TOFPET2 ASIC and SiPMs from various producers. DOI and TOF performance are compared to results using custom-made NINO 32-chip based electronics.&#xD;Main Results: Using a Broadcom NUV-MT array, the detector module read out by the TOFPET2 ASIC demonstrates a DOI resolution of 2.6 ± 0.2 mm full width at half maximum (FWHM) and a coincidence time resolution (CTR) of 216 ± 6 ps FWHM. When read out using the NINO 32-chip based electronics, the same module achieves a DOI resolution of 2.5 ± 0.2 mm and a CTR of 170 ± 5 ps.&#xD;Significance: The prototype module, read out by commercial electronics and using state-of-the-art SiPMs, achieves a DOI performance comparable to that obtained with custom-made electronics and a CTR of around 200 ps. This approach is scalable to thousands of channels, with only a deterioration in timing resolution compared to the custom-made electronics, which achieve a CTR of 140 ps using a standard non-DOI module.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142865061","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A comparative study of experimental and simulated ultrasound beam propagation through cranial bones.","authors":"Alisa Krokhmal, Ian Simcock, Bradley E Treeby, Eleanor Martin","doi":"10.1088/1361-6560/ada19d","DOIUrl":"https://doi.org/10.1088/1361-6560/ada19d","url":null,"abstract":"<p><strong>Objective: </strong>Transcranial ultrasound is used in a variety of treatments, including neuromodulation, opening the blood-brain barrier (BBB), and high intensity focused ultrasound (HIFU) therapies. To ensure safety and efficacy of these treatments, numerical simulations of the ultrasound field within the brain are used for treatment planning and evaluation. This study investigates the accuracy of numerical modelling of the propagation of focused ultrasound through cranial bones. &#xD;&#xD;Approach: Holograms of acoustic fields after propagation through four human skull specimens were measured for frequencies ranging from 270 kHz to 1 MHz, using both quasi-continuous and pulsed modes. The open-source k-Wave toolbox was employed for simulations, using an equivalent-source hologram and a uniform bowl source with parameters that best matched the measured free-field pressure distribution. &#xD;&#xD;Main results: The average absolute error in k-Wave simulations with sound speed and density derived from CT scans compared to measurements was 15% for the spatial-peak acoustic pressure amplitude, 2.7 mm for the position of the focus, and 35% for the focal volume. Optimised uniform bowl sources achieved calculation accuracy comparable to that of the hologram sources. &#xD;&#xD;Significance: This method is demonstrated as a suitable tool for prediction of focal position, size and overall distribution of transcranial ultrasound fields. The accuracy of the shape and position of the focal region demonstrate the suitability of the sound speed and density mapping used here. However, large errors in pressure amplitude and transmission loss in some individual cases show that alternative methods for mapping individual skull attenuation are needed and the possibility of considerable errors in pressure amplitude should be taken into account when planning focused ultrasound studies or interventions in the human brain, and appropriate safety margins should be used.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142865004","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The textures of sarcoidosis: quantifying lung disease through variograms.","authors":"William Lippitt, Lisa A Maier, Tasha E Fingerlin, David A Lynch, Ruchi Yadav, Jared Rieck, Andrew C Hill, Shu-Yi Liao, Margaret M Mroz, Briana Q Barkes, Kum Ju Chae, Hye Jeon Hwang, Nichole E Carlson","doi":"10.1088/1361-6560/ada19c","DOIUrl":"https://doi.org/10.1088/1361-6560/ada19c","url":null,"abstract":"<p><strong>Objective: </strong>Sarcoidosis is a granulomatous disease affecting the lungs in over 90% of patients. Qualitative assessment of chest CT by radiologists is standard clinical practice and reliable quantification of disease from CT would support ongoing efforts to identify sarcoidosis phenotypes. Standard imaging feature engineering techniques such as radiomics suffer from extreme sensitivity to image acquisition and processing, potentially impeding generalizability of research to clinical populations. In this work, we instead investigate approaches to engineering variogram-based features with the intent to identify a robust, generalizable pipeline for image quantification in the study of sarcoidosis. &#xD;&#xD;Approach: For a cohort of more than 300 individuals with sarcoidosis, we investigated 24 feature engineering pipelines differing by decisions for image registration to a template lung, empirical and model variogram estimation methods, and feature harmonization for CT scanner model, and subsequently 48 sets of phenotypes produced through unsupervised clustering. We then assessed sensitivity of engineered features, phenotypes produced through unsupervised clustering, and sarcoidosis disease signal strength to pipeline. &#xD;&#xD;Main results: We found that variogram features had low to mild association with scanner model and associations were reduced by image registration. For each feature type, features were also typically robust to all pipeline decisions except image registration. Strength of disease signal as measured by association with pulmonary function testing and some radiologist visual assessments was strong (optimistic AUC $approx 0.9$, $pll0.0001$ in models for architectural distortion, conglomerate mass, fibrotic abnormality, and traction bronchiectasis) and fairly consistent across engineering approaches regardless of registration and harmonization for CT scanner. &#xD;&#xD;Significance: Variogram-based features appear to be a suitable approach to image quantification in support of generalizable research in pulmonary sarcoidosis.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142865067","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Feasibility of reconstructing in-vivo patient 3D dose distributions from 2D EPID image data using convolutional neural networks.","authors":"Ning Gao, Bo Cheng, Zhi Wang, Didi Li, Yankui Chang, Qiang Ren, Xi Pei, Chengyu Shi, X George Xu","doi":"10.1088/1361-6560/ada19b","DOIUrl":"https://doi.org/10.1088/1361-6560/ada19b","url":null,"abstract":"<p><strong>Objective: </strong>The primary purpose of this work is to demonstrate the feasibility of a deep convolutional neural network (dCNN) based algorithm that uses two-dimensional (2D) EPID images and CT images as input to reconstruct 3D dose distributions inside the patient. &#xD;&#xD;Approach: To generalize dCNN training and testing data, geometric and materials models of a VitalBeam accelerator treatment head and a corresponding EPID imager were constructed in detail in the GPU-accelerated Monte Carlo dose computing software, ARCHER. The EPID imager pixel spatial resolution ranging from 1.0 mm to 8.5 mm was studied to select optimal pixel size for simulation. For purposes of training the U-Net-based dCNN, a total of 101 clinical IMRT cases - 81 for training, 10 for validation, and 10 for testing - were simulated to produce comparative data of 3D dose distribution versus 2D EPID image data. The model's accuracy was evaluated by comparing its predictions with Monte Carlo dose.&#xD;&#xD;Main Results: Using the optimal EPID pixel size of 1.5 mm, it took about 18 min to simulate the particle transport in patient-specific CT and EPID imager per a single field. In contrast, the trained dCNN can predict 3D dose distributions in about 0.35s. The average 3D gamma passing rates between ARCHER and predicted doses are 99.02±0.57% (3%/3mm) and 96.85±1.22% (2%/2 mm) for accumulated fields, respectively. DVH data suggest that the proposed dCNN 3D dose prediction algorithm is accurate in evaluating treatment goals.&#xD;&#xD;Significance: This study has proposed a novel deep-learning model that is accurate and rapid in predicting 3D patient dose from 2D EPID images. The computational speed is expected to facilitate clinical practice for EPID-based in-vivo patient-specific quality assurance towards adaptive radiation therapy.&#xD.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142865064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computationally efficient collimator-detector response compensation in high energy SPECT using 1D convolutions and rotations.","authors":"Lucas Alexander Polson, Pedro Esquinas, Sara Kurkowska, Chenguang Li, Peyman Sheikhzadeh, Mehrshad Abbasi, Saeed Farzenefar, Seyyede Mirabedian, Carlos Felipe Uribe, Arman Rahmim","doi":"10.1088/1361-6560/ada10a","DOIUrl":"https://doi.org/10.1088/1361-6560/ada10a","url":null,"abstract":"<p><strong>Objective: </strong>Modeling of the collimator-detector response (CDR) in SPECT reconstruction enables improved resolution and accuracy, and is thus important for quantitative imaging applications such as dosimetry. The implementation of CDR modeling, however, can become a computational bottleneck when there are substantial components of septal penetration and scatter in the acquired data, since a direct convolution-based approach requires large 2D kernels. This work proposes a 1D convolution and rotation-based CDR model that reduces reconstruction times but maintains consistency with models that employ 2D convolutions. To enable open-source development and use of these models in image reconstruction, we release a SPECTPSFToolbox repository for the PyTomography project on GitHub.</p><p><strong>Approach: </strong>A 1D/rotation-based CDR model was formulated and subsequently fit to Monte Carlo point source data representative of Lu-177, I-131, and Ac-225 imaging. Computation times of (i) the proposed 1D/rotation-based model and (ii) a traditional model that uses 2D convolutions were compared for typical SPECT matrix sizes. Both CDR models were then used in the reconstruction of Monte Carlo, physical phantom, and patient data; the models were compared by quantifying total counts in hot regions of interest (ROIs) and activity contrast between hot ROIs and background regions.</p><p><strong>Results: </strong>For typical matrix sizes in SPECT reconstruction, application of the 1D/rotation-based model provides a two-fold computational speed-up over the 2D model when running on GPU. Only small differences between the 1D/rotation-based and 2D models (order of 1%) were obtained for count and contrast quantification in select ROIs.</p><p><strong>Significance: </strong>A technique for CDR modeling in SPECT was proposed that (i) significantly speeds up reconstruction times, and (ii) yields nearly identical reconstructions to traditional 2D convolution based CDR techniques. The released toolbox will permit open-source development of similar models for different isotopes and collimators.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142854488","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Path planning for flexible needle based on both insertion mechanism kinematics and needle bending model.","authors":"Yan-Jiang Zhao, He Zhang, Haiyan Du, Cheng Qian, Ye-Xin Jin, Yin Chen, Guang-Zhe Yang","doi":"10.1088/1361-6560/ada0a1","DOIUrl":"https://doi.org/10.1088/1361-6560/ada0a1","url":null,"abstract":"<p><strong>Objective: </strong>Due to the limited operating space in the magnetic resonance (MR) environment, there is coupled motion in the insertion mechanism, which not only reduces the flexibility of the robot but also challenges the insertion path planning. Meanwhile, the path planning is also restricted by the bending rule of the flexible needle, thus the bending model of the needle is also essentially built.</p><p><strong>Approach: </strong>This paper proposes a path planner for the flexible needle based on both the coupled motion kinematics of the insertion robot and the bending model of the flexible needle. A kinematic analysis for the coupled motion of insertion robot is performed. And the bending model of flexible needle is established based on the needle-tissue interactions. The position and posture of the needle insertion at the entry point are obtained by the calculation of the target position and the analysis of the bending model. And the rotation or translation coordinates of each robot joint are calculated by the inverse kinematics of the insertion robot. Then the path planning based on the coupled kinematics and the bending model is realized.</p><p><strong>Main results: </strong>The insertion experiments were performed for each target of G1 and G2. The root mean square errors were 0.83 mm and 0.74 mm, and the maximum errors were 1.1 mm and 0.9 mm for G1 and G2, respectively. The experimental results show that the effectiveness and accuracy of the path planning can meet the requirements for a general minimally invasive surgery, so the proposed path planning algorithm is feasible.</p><p><strong>Significance: </strong>This study provides a new solution for the path planning of insertion robots for the minimally invasive surgery. This method can meet the insertion mechanism working within the limited operating space in the MR environment and has a high application value in future clinical medicine.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142847464","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bimodal PET/MRI generative reconstruction based on VAE architectures.","authors":"V Gautier, A Bousse, F Sureau, C Comtat, V Maxim, B Sixou","doi":"10.1088/1361-6560/ad9133","DOIUrl":"10.1088/1361-6560/ad9133","url":null,"abstract":"<p><p><i>Objective.</i>In this study, we explore positron emission tomography (PET)/magnetic resonance imaging (MRI) joint reconstruction within a deep learning framework, introducing a novel synergistic method.<i>Approach.</i>We propose a new approach based on a variational autoencoder (VAE) constraint combined with the alternating direction method of multipliers (ADMM) optimization technique. We explore three VAE architectures, joint VAE, product of experts-VAE and multimodal JS divergence (MMJSD), to determine the optimal latent representation for the two modalities. We then trained and evaluated the architectures on a brain PET/MRI dataset.<i>Main results.</i>We showed that our approach takes advantage of each modality sharing information to each other, which results in improved peak signal-to-noise ratio and structural similarity as compared with traditional reconstruction, particularly for short acquisition times. We find that the one particular architecture, MMJSD, is the most effective for our methodology.<i>Significance.</i>The proposed method outperforms conventional approaches especially in noisy and undersampled conditions by making use of the two modalities together to compensate for the missing information.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142626237","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Denoising proton reference dosimetry spectrum using a large area ionization chamber - physical basis and type A uncertainty.","authors":"Hong Qi Tan, Kah Seng Lew, Calvin Wei Yang Koh, Kang Hao Lee, Clifford Ghee Ann Chua, Andrew Wibawa, Zubin Master, James Cheow Lei Lee, Sung Yong Park","doi":"10.1088/1361-6560/ada085","DOIUrl":"https://doi.org/10.1088/1361-6560/ada085","url":null,"abstract":"<p><strong>Objective: </strong>&#xD;Reference dosimetry measurement in a pencil beam scanning system can exhibit dose fluctuation due to intra-spill spot positional drift. This results in a noisy reference dosimetry measurement against energy which could introduce errors in monitor unit calibration. The aim of this study is to investigate the impact of smoothing the reference dosimetry measurements on the type A uncertainty.&#xD;Methods:&#xD;The reference dosimetry measurement (D_w/MU) with a PTW 34045 advanced Markus chamber placed at 2 cm depth and a 10 x 10 cm^2 scanned field are performed for 98 energy layers on five non-consecutive days using a water tank. The PTW 34089 large area ionization chamber (LAIC) is placed at the same depth and the charges are measured with a single spot irradiation (M_spot^LAIC). (D_w/MU) and M_spot^LAIC are fitted with a linear and quadratic function to obtain a smooth plot of (D_w/MU) against the proton energy (reference dosimetry curve). Type A uncertainty of the measured reference dosimetry curve is compared against the de-noised fitted curve. &#xD;Results:&#xD;The repeatability of reference dosimetry measurement shows relative difference of up to 2.3% across the five days. The linear and quadratic fits between LAIC charges and the (D_w/MU) from PTW 34045 show a high R^2 values of more than 0.95. The maximum type A uncertainty of the de-noised reference dosimetry curve is lower (0.69% at 70.2 MeV) compared to the measured one (0.88% at 77.5 MeV). However, the average type A uncertainty of the denoised curve across all energies is higher compared to the measurements (0.50% versus 0.43%).&#xD;Conclusion:&#xD;We have presented the physical basis and procedure for fitting the charges measured with a LAIC to the reference dosimetry curve. The fitted reference dosimetry curve avoids large error in any energy layer but increases the average type A uncertainty across energies and should be used with caution. &#xD;&#xD;&#xD.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142847462","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"RADD-CycleGAN: unsupervised reconstruction of high-quality ultrasound image based on CycleGAN with residual attention and dual-domain discrimination.","authors":"Mateng Si, Musheng Wu, Qing Wang","doi":"10.1088/1361-6560/ad997f","DOIUrl":"10.1088/1361-6560/ad997f","url":null,"abstract":"<p><p>Plane wave (PW) imaging is fast, but limited by poor imaging quality. Coherent PW compounding (CPWC) improves image quality but decrease frame rate. In this study, we propose a modified CycleGAN model that combines a residual attention module with a space-frequency dual-domain discriminator, termed RADD-CycleGAN, to rapidly reconstruct high-quality ultrasound images. To enhance the ability to reconstruct image details, we specially design a process of hybrid dynamic and static channel selection followed by the frequency domain discriminator. The low-quality images are generated by the 3-angle CPWC, while the high-quality images are generated as real images (ground truth) by the 75-angle CPWC. The training set includes unpaired images, whereas the images in the test set are paired to verify the validity and superiority of the proposed model. Finally, we respectively design ablation and comparison experiments to evaluate the model performance. Compared with the basic CycleGAN, our proposed method reaches a better performance, with a 7.8% increase in the peak signal-to-noise ratio and a 22.2% increase in the structural similarity index measure. The experimental results show that our method achieves the best unsupervised reconstruction from low quality images in comparison with several state-of-the-art methods.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.3,"publicationDate":"2024-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142771336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}