Article
Radiology, Nuclear Medicine & Medical Imaging
Emil Y. Sidky, Xiaochuan Pan
Summary: The purpose of this challenge is to find a deep-learning technique that can achieve the minimum root mean square error for sparse-view CT image reconstruction under ideal conditions. The challenge involves a 2D breast CT simulation and a large training set to train the networks. About 60 groups participated in the challenge, achieving significant improvement in reconstruction accuracy.
Article
Nutrition & Dietetics
Yoon Seong Lee, Namki Hong, Joseph Nathanael Witanto, Ye Ra Choi, Junghoan Park, Pierre Decazes, Florian Eude, Chang Oh Kim, Hyeon Chang Kim, Jin Mo Goo, Yumie Rhee, Soon Ho Yoon
Summary: The study developed and validated a deep neural network model for automatic volumetric segmentation of body composition on whole-body CT images, presenting potential applications in sarcopenia assessment and metabolic evaluation of whole-body muscle and fat tissues.
CLINICAL NUTRITION
(2021)
Article
Automation & Control Systems
Wenqiang Li, Yuk Ming Tang, Ziyang Wang, Kai Ming Yu, Suet To
Summary: Automatic vertebrae segmentation using CT plays a crucial role in automated spine analysis, and recent advancements in deep learning have led to precise performance through deep convolutional neural networks. While DCNN-based semantic segmentation algorithms have advantages, they face limitations that are addressed by the proposed novel algorithm, which includes encoder-decoder framework, Layer Normalization, Atrous Residual Path, and a 3D Attention Module to improve segmentation accuracy. Experimental results show competitive performance compared to existing methods for automatic vertebrae semantic segmentation.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2022)
Article
Computer Science, Interdisciplinary Applications
Xi Tao, Yongbo Wang, Liyan Lin, Zixuan Hong, Jianhua Ma
Summary: The study trains DNN to reconstruct CT images from VVBP-Tensor, providing lossless information and preserving fine details. The learning strategy can be seen as a generalization of conventional methods and potentially inspire algorithm development. The VVBP-Tensor domain learning framework shows significant improvement over traditional image and projection-based learning frameworks.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Andreas Heinrich, Sebastian Schenkl, David Buckreus, Felix V. Guettler, Ulf K-M. Teichgraeber
Summary: This study evaluated the sensitivity of CT-based thermometry for fat, muscle, and bone tissues, finding that VMI provides higher temperature sensitivity for these tissues. The reconstruction algorithms ASIR-V and DLIR do not significantly impact CT-based thermometry.
EUROPEAN RADIOLOGY
(2022)
Article
Engineering, Biomedical
Dan Lior, Charles Puelz, Colin Edwards, Silvana Molossi, Boyce E. Griffith, Ravi K. Birla, Craig G. Rusin
Summary: This paper presents a semi-automatic method for constructing volumetric models of the aortic valve using computed tomography angiography images. The method uses manually selected samples of the aortic segmentation derived from the images to inform the model construction. Valve models for pediatric patients are created and simulation results show that the method produces functional valves that generate pressure and flow waveforms similar to clinical observations.
ANNALS OF BIOMEDICAL ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Lianying Chao, Zhiwei Wang, Haobo Zhang, Wenting Xu, Peng Zhang, Qiang Li
Summary: Cone beam computed tomography (CBCT) is widely used in image-guided surgery and radiotherapy, but it poses risks of ionizing radiation to patients. Researchers have developed a dual convolutional neural network architecture (DualCNN) to eliminate streak artifacts in sparse-view CBCT images, and experimental results demonstrate its significantly higher performance compared to other methods.
Article
Computer Science, Interdisciplinary Applications
Yixing Huang, Alexander Preuhs, Michael Manhart, Guenter Lauritsch, Andreas Maier
Summary: Data truncation in CT reconstruction causes artifacts and missing anatomical structures. Deep learning has shown impressive results in CT reconstruction, but concerns remain about its robustness in clinical applications. A plug-and-play method is proposed for truncation correction, integrating deep learning and conventional algorithms for better robustness and interpretability. Demonstration on state-of-the-art deep learning methods shows the efficacy of the proposed method in improving image quality and reducing errors in noisy cases.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2021)
Article
Computer Science, Interdisciplinary Applications
Weiwen Wu, Dianlin Hu, Chuang Niu, Hengyong Yu, Varut Vardhanabhuti, Ge Wang
Summary: The article introduces a Dual-domain Residual-based Optimization NEtwork (DRONE) to address the challenge of sparse-view CT reconstruction. The network consists of three modules for embedding, refinement, and awareness, aimed at suppressing sparse-view artifacts and recovering image details to optimize image quality.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Shota Masuda, Yoshitake Yamada, Kazuya Minamishima, Yoshiki Owaki, Akihisa Yamazaki, Masahiro Jinzaki
Summary: This study assessed the effects of deep learning image reconstruction (DLIR) and hybrid iterative reconstruction (HIR) on the image quality of virtual monochromatic spectral (VMS) images. The results showed that DLIR achieved better noise reduction than HIR. Additionally, the image quality of VMS-DLIR and 120 kVp-DLIR may decrease in medium contrast tasks and increase in high contrast tasks.
EUROPEAN JOURNAL OF RADIOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Jonas Teuwen, Nikita Moriakov, Christian Fedon, Marco Caballo, Ingrid Reiser, Pedrag Bakic, Eloy Garcia, Oliver Diaz, Koen Michielsen, Ioannis Sechopoulos
Summary: This study introduces a deep learning-based reconstruction algorithm for digital breast tomosynthesis (DBT), named DBToR, which is capable of estimating true breast density and patient-specific radiation dose with high accuracy through model training and testing.
MEDICAL IMAGE ANALYSIS
(2021)
Article
Engineering, Aerospace
Fuhao Zhang, Weixuan Zhang, Qingchun Lei, Xuesong Li, Yuyang Li, Min Xu
Summary: Volumetric reconstructions of transparent or translucent mediums are critical for various applications. This study proposes a neural volume reconstruction technique (NVRT) that uses a neural network to represent the continuous flame luminosity implicitly. The NVRT method is superior to the traditional algebraic reconstruction technique (ART) in terms of reconstruction fidelity, resistance to noise, and computational cost.
AEROSPACE SCIENCE AND TECHNOLOGY
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Yazdan Salimi, Isaac Shiri, Azadeh Akavanallaf, Zahra Mansouri, Hossein Arabi, Habib Zaidi
Summary: This study aimed to improve patient positioning accuracy using a CT localizer and a deep neural network. The performance of the model was evaluated by comparing the estimated body centerline with the ground truth. The proposed method showed comparable accuracy to alternative methods.
EUROPEAN RADIOLOGY
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Mineka Sato, Yasutaka Ichikawa, Kensuke Domae, Kazuya Yoshikawa, Yoshinori Kanii, Akio Yamazaki, Naoki Nagasawa, Motonori Nagata, Masaki Ishida, Hajime Sakuma
Summary: Compared to hybrid iterative reconstruction (IR), deep learning image reconstruction (DLIR) improves vessel conspicuity, contrast-to-noise ratio (CNR), and lesion conspicuity of virtual monochromatic and iodine density images in abdominal contrast-enhanced dual-energy computed tomography (DECT).
EUROPEAN RADIOLOGY
(2022)
Article
Engineering, Electrical & Electronic
Yuan Gao, Hui Tang, Rongjun Ge, Jin Liu, Xin Chen, Yan Xi, Xu Ji, Huazhong Shu, Jian Zhu, Gouenou Coatrieux, Jean-Louis Coatrieux, Yang Chen
Summary: Orthopedic spine disease is a common condition that often requires accurate diagnosis using orthopedic computed tomography (CT) image data. However, in some situations where 3-D imaging equipment is lacking or time is limited, a method based on 2-D x-ray images is needed. In this study, a novel 3-D spine reconstruction technique called 3DSRNet is proposed. It utilizes a generative adversarial network (GAN) architecture and novel modules to achieve accurate and efficient reconstruction by integrating local bone surface information, long-range relation spinal structure information, and spine texture features.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Computer Science, Interdisciplinary Applications
Rui Yan, Liangqiong Qu, Qingyue Wei, Shih-Cheng Huang, Liyue Shen, Daniel L. Rubin, Lei Xing, Yuyin Zhou
Summary: In this paper, a robust and label-efficient self-supervised federated learning (FL) framework for medical image analysis is proposed. The method introduces a novel Transformer-based self-supervised pre-training paradigm, which enhances the robustness of models against heterogeneous data and improves knowledge transfer to downstream models. Extensive experiments on simulated and real-world medical imaging non-IID federated datasets demonstrate the effectiveness of the proposed method.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Lianli Liu, Liyue Shen, Yong Yang, Emil Schueler, Wei Zhao, Gordon Wetzstein, Lei Xing
Summary: In this study, a NeRP beam data modeling technique was developed to predict beam characteristics from sparse measurements and detect data collection errors. The model showed promising results in predicting beam data and verifying its accuracy, which can simplify the commissioning and QA process without compromising the quality of medical physics service.
Article
Radiology, Nuclear Medicine & Medical Imaging
Oscar Pastor-Serrano, Peng Dong, Charles Huang, Xing Lei, Zoltan Perko
Summary: This study presents a deep learning algorithm that accurately predicts broad photon beam dose distributions in a few milliseconds by combining transformer and convolutional layers. The proposed algorithm maps patient geometries and beam information to corresponding dose distributions, achieving high accuracy and speed.
Article
Computer Science, Artificial Intelligence
Varun Vasudevan, Maxime Bassenne, Md Tauhidul Islam, Lei Xing
Summary: Previous studies on image classification using graph neural networks (GNNs) have mainly focused on graphs generated from a regular grid of pixels or similar-sized superpixels. However, this study investigates image classification using graphs generated from an image-specific number of multiscale superpixels. The proposed WaveMesh algorithm calculates the number and sizes of superpixels in an image based on its content, resulting in structurally different superpixel graphs compared to similar-sized superpixels. Experimental results show that the SplineCNN network learns from multiscale WaveMesh superpixels on-par with similar-sized superpixels, while poor cluster assignment negatively affects the network's performance.
PATTERN RECOGNITION LETTERS
(2023)
Article
Engineering, Biomedical
Rongjun Ge, Fanqi Shi, Yang Chen, Shujun Tang, Hailong Zhang, Xiaojian Lou, Wei Zhao, Gouenou Coatrieux, Dazhi Gao, Shuo Li, Xiaoli Mai
Summary: In this study, an Average Super-Resolution Generative Adversarial Network (ASRGAN) is proposed to recover inter-slice information from CT images with different quality. Experimental results show that ASRGAN outperforms other methods in reconstruction, with a 2.42 db improvement in PSNR. Additionally, based on its reconstruction results, it further improves 3D segmentation of abdominal tumors and the pancreas.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Computer Science, Interdisciplinary Applications
Xinpeng Ding, Xinjian Yan, Zixun Wang, Wei Zhao, Jian Zhuang, Xiaowei Xu, Xiaomeng Li
Summary: This paper introduces a timestamp supervision method for surgical phase recognition, which reduces the manual annotation cost. The proposed UATD method generates reliable pseudo labels for training and achieves competitive results in surgical phase recognition.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Multidisciplinary Sciences
Md Tauhidul Islam, Lei Xing
Summary: The authors develop a cartography strategy based on gene-gene interactions to transform high-dimensional gene expression data into a spatially configured genomap, enabling accurate deep pattern discovery. This approach presents significant challenges and opportunities in the field of single cell genomics for biomedical research. The unique cartography method captures gene interactions in the spatial configuration of genomaps, allowing for the extraction of deep genomic interaction features and the discovery of discriminative patterns in the data.
NATURE COMMUNICATIONS
(2023)
Article
Engineering, Biomedical
Xiao Jia, Eugene Shkolyar, Mark A. Laurie, Okyaz Eminaga, Joseph C. Liao, Lei Xing
Summary: This study developed a deep learning algorithm for accurate detection of bladder tumors in white-light cystoscopy. The algorithm, CystoNet-T, uses a transformer-augmented pyramidal CNN architecture to improve tumor detection performance. The algorithm achieved high F1 and AP scores on a test set of patient data.
PHYSICS IN MEDICINE AND BIOLOGY
(2023)
Article
Multidisciplinary Sciences
Yuming Jiang, Zhicheng Zhang, Wei Wang, Weicai Huang, Chuanli Chen, Sujuan Xi, M. Usman Ahmad, Yulan Ren, Shengtian Sang, Jingjing Xie, Jen-Yeu Wang, Wenjun Xiong, Tuanjie Li, Zhen Han, Qingyu Yuan, Yikai Xu, Lei Xing, George A. Poultsides, Guoxin Li, Ruijiang Li
Summary: Significant progress has been made in using deep learning for cancer detection and diagnosis, but predicting treatment response and outcomes remains challenging. This study presents a biology-guided deep learning approach that simultaneously predicts tumor microenvironment status and treatment outcomes from medical images. The model was validated for gastric cancer prognosis and response to adjuvant chemotherapy, and it complements clinically approved biomarkers, providing information for patient selection.
NATURE COMMUNICATIONS
(2023)
Proceedings Paper
Optics
Xiao Jia, Eugene Shkolyar, Okyaz Eminaga, Mark A. Laurie, Zixia Zhou, Timothy Lee, Md Tauhidul Islam, Max Q. -H. Meng, Joseph C. Liao, Lei Xing
Summary: In this study, a deep-learning algorithm called CystoNet-F was developed to enhance the detection of flat lesions in white-light cystoscopy. By incorporating domain translation, transfer learning, and region of interest detection, the algorithm improves the ability to detect lesions in standard white-light cystoscopy.
ADVANCED PHOTONICS IN UROLOGY 2023
(2023)
Proceedings Paper
Optics
Mark Laurie, Okyaz Eminaga, Eugene Shkolyar, Xiao Jia, Timothy Lee, Jin Long, Md Tauhidul Islam, Hubert Lau, Lei Xing, Joseph C. Liao
Summary: In this study, four advanced sequential models (SlowFast, Multiscale Vision Transformers, X3D, and CNN-LSTM) were used to train and classify bladder tumors. The best performing model was X3D with a sequence length of 8, achieving 100% sensitivity, 94.7% sensitivity, and 80.0% specificity. Sequential modeling has the potential to accurately classify a wide variety of bladder tumors in a real-time setting.
ADVANCED PHOTONICS IN UROLOGY 2023
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Bowen Song, Liyue Shen, Lei Xing
Summary: Recently, deep learning has been used to address medical image reconstruction problems in CT scans with limited views. However, these models have difficulty generalizing to new domains and the privacy concerns limit access to training data. To overcome these challenges, a source-free black-box test-time adaptation method called PINER is introduced. By leveraging implicit neural representation learning, the method can adapt to unknown noise levels and outperforms state-of-the-art algorithms.
2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV)
(2023)
Article
Computer Science, Artificial Intelligence
Shengtian Sang, Yuyin Zhou, Md Tauhidul Islam, Lei Xing
Summary: This article introduces a novel attention-based method called AFMA, which addresses the issue of information loss in small object segmentation. By utilizing different feature levels of the original image, AFMA quantifies the inner relationship between large and small objects of the same category, compensating for the loss of high-level feature information and improving segmentation accuracy. Extensive experiments demonstrate the substantial and consistent improvement of small object segmentation achieved by our method.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Zixia Zhou, Md Tauhidul Islam, Lei Xing
Summary: Deep learning-based diagnosis is essential in modern healthcare. An optimal design of deep neural networks (DNNs) is crucial for high-performance diagnosis. Existing supervised DNNs based on convolutional layers tend to have limited feature exploration ability, which compromises the network performance.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Lianli Liu, Liyue Shen, Adam Johansson, James M. Balter, Yue Cao, Lucas Vitzthum, Lei Xing
Summary: A method for volumetric MRI reconstruction using a NeRP model has been proposed, training the model with only one inhale and one exhale 3D MRI to learn prior information of patient breathing motion for sparse image reconstruction. The method has the potential to support 3D motion tracking during MR-guided radiotherapy and promises a major simplification of the workflow.
