Real Computed Tomography Images Research Articles

SeRL: Style-embedding representation learning for unsupervised CT images synthesis from unpaired MR images

Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related deaths worldwide, and the fastest-growing cause of cancer deaths in the United States. Computed tomography (CT) and magnetic resonance (MR) imaging are the key image-based examination methods for HCC patients. Patients may alternatively undergo CT and MR exams given a radiation dose of the former and the cost of the latter. Due to organ representation differences between CT and MR, it’s a great challenge to detect the tumors from a time series of data consisting of both CT and MR by a single model. The annotations for these time series data will consume much time and labor for the physicians. Thus, we propose our style-embedding representation learning (SeRL) for unsupervised and unpaired abdomen CT and MR translation. Different from current medical image translation models, the style-representation information from real CT and real MR images has been embedded in the translation process to bypass some local minima during the convergence process and improve the synthesis results. Our patch-corrosion augmentation method enhances the style-embedding representation learning by bringing more diversity to the training data. Combined with the self-attention module, SeRL eliminates the noise caused by low grayscale pixel values during translation. Results on the unpaired HCC patient's CT and MR images show that our proposed SeRL is able to generate high quality CT images from MR ones. Evaluations such as Frechet Inception Distance (FID), Sliced Wasserstein Distance (SWD), and liver segmentation dice score are utilized to demonstrate our advantages over other state-of-the-art unsupervised methods.

A nonlinear total variation based computed tomography (CT) image reconstruction method using gradient reinforcement.

Compressed sensing-based reconstruction algorithms have been proven to be more successful than analytical or iterative methods for sparse computed tomography (CT) imaging by narrowing down the solution set thanks to its ability to seek a sparser solution. Total variation (TV), one of the most popular sparsifiers, exploits spatial continuity of features by restricting variation between two neighboring pixels in each direction as using partial derivatives. When the number of projections is much fewer than the one in conventional CT, which results in much less sampling rate than the minimum required one, TV may not provide satisfactory results. In this study, a new regularizer is proposed which seeks for a sparser solution by reinforcing the gradient of TV and empowering the spatial continuity of features. The experiments are done by using both analitical phantom and real human CT images and the results are compared with conventional, four-directional, and directional TV algorithms by using contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR) and Structural Similarity Index (SSIM) metrics. Both quantitative and visual evaluations show that the proposed method is promising for sparse CT image reconstruction by reducing the background noise while preserving the features and edges.

Industrial computed tomography for three-dimensional cultural relic model reconstruction based on L1-αL2+TV norm minimization

Industrial computed tomography (ICT) has emerged as a valuable tool for the conservation of cultural relics since it enables the extraction of information about the internal structure of artifacts without causing any damage. However, sparse and finite angle projection computed tomography (CT) often generates images with a large number of artifacts caused by insufficient data. This paper introduces a new cone-beam CT hybrid model based on compressed sensing (CS) theory and total variance (TV) denoising using the Split Bregman algorithm to address this problem, for the first time using compressed sensing theory for CT reconstruction of cultural relics. The model combines the L1 and L2 norms and optimizes the total variance of the image as the objective function. To enhance the reconstruction of L1-L2, the paper employs the continuation strategy of α and proposes a new combination function to update α. Consequently, the reconstruction results are compared with those of traditional common methods, and the effect of different α update functions is observed on the reconstruction results. The results from the model and real CT images demonstrate that the proposed L1-αL2 + TV method yields superior reconstruction results compared to ART + TV, SART, and L1/L2 algorithms based on objective evaluation indexes such as RMSE, SSIM, and PSNR, the proposed L1-αL2 + TV method achieves SSIM improvement of at least 15.2 %. Moreover, the newly proposed combined function is better at preserving the grain of the image, resulting in a much-improved reconstruction of the contours, making it highly beneficial for subsequent conservation and restoration of heritage. Overall, this proposed L1-αL2 + TV mixture model provides low-noise and high-quality reconstruction results, while maintaining the grain of the original image, making it a great boon for heritage conservation and restoration.

Development, validation, and simplification of a scanner-specific CT simulator.

Simulated computed tomography (CT) images allow for knowledge of the underlying ground truth and for easy variation of imaging conditions, making them ideal for testing and optimization of new applications or algorithms. However, simulating all processes that affect CT images can result in simulations that are demanding in terms of processing time and computer memory. Therefore, it is of interest to determine how much the simulation can be simplified while still achieving realistic results. To develop a scanner-specific CT simulation using physics-based simulations for the position-dependent effects and shift-invariant image corruption methods for the detector effects. And to investigate the impact on image realism of introducing simplifications in the simulation process that lead to faster and less memory-demanding simulations. To make the simulator realistic and scanner-specific, the spatial resolution and noise characteristics, and the exposure-to-detector output relationship of a clinical CT system were determined. The simulator includes a finite focal spot size, raytracing of the digital phantom, gantry rotation during projection acquisition, and finite detector element size. Previously published spectral models were used to model the spectrum for the given tube voltage. The integrated energy at each element of the detector was calculated using the Beer-Lambert law. The resulting angular projections were subsequently corrupted by the detector modulation transfer function (MTF), and by addition of noise according to the noise power spectrum (NPS) and signal mean-variance relationship, which were measured for different scanner settings. The simulated sinograms were reconstructed on the clinical CT system and compared to real CT images in terms of CT numbers, noise magnitude using the standard deviation, noise frequency content using the NPS, and spatial resolution using the MTF throughout the field of view (FOV). The CT numbers were validated using a multi-energy CT phantom, the noise magnitude and frequency were validated with a water phantom, and the spatial resolution was validated with a tungsten wire. These metrics were compared at multiple scanner settings, and locations in the FOV. Once validated, the simulation was simplified by reducing the level of subsampling of the focal spot area, rotation and of detector pixel size, and the changes in MTFs were analyzed. The average relative errors for spatial resolution within and across image slices, noise magnitude, and noise frequency content within and across slices were 3.4%, 3.3%, 4.9%, 3.9%, and 6.2%, respectively. The average absolute difference in CT numbers was 10.2HU and the maximum was 22.5HU. The simulation simplification showed that all subsampling can be avoided, except for angular, while the error in frequency at 10% MTF would be maximum 16.3%. The simulation of a scanner-specific CT allows for the generation of realistic CT images by combining physics-based simulations for the position-dependent effects and image-corruption methods for the shift-invariant ones. Together with the available ground truth of the digital phantom, it results in a useful tool to perform quantitative analysis of reconstruction or post-processing algorithms. Some simulation simplifications allow for reduced time and computer power requirements with minimal loss of realism.

Decomposition of musculoskeletal structures from radiographs using an improved CycleGAN framework

This paper presents methods of decomposition of musculoskeletal structures from radiographs into multiple individual muscle and bone structures. While existing solutions require dual-energy scan for the training dataset and are mainly applied to structures with high-intensity contrast, such as bones, we focused on multiple superimposed muscles with subtle contrast in addition to bones. The decomposition problem is formulated as an image translation problem between (1) a real X-ray image and (2) multiple digitally reconstructed radiographs, each of which contains a single muscle or bone structure, and solved using unpaired training based on the CycleGAN framework. The training dataset was created via automatic computed tomography (CT) segmentation of muscle/bone regions and virtually projecting them with geometric parameters similar to the real X-ray images. Two additional features were incorporated into the CycleGAN framework to achieve a high-resolution and accurate decomposition: hierarchical learning and reconstruction loss with the gradient correlation similarity metric. Furthermore, we introduced a new diagnostic metric for muscle asymmetry directly measured from a plain X-ray image to validate the proposed method. Our simulation and real-image experiments using real X-ray and CT images of 475 patients with hip diseases suggested that each additional feature significantly enhanced the decomposition accuracy. The experiments also evaluated the accuracy of muscle volume ratio measurement, which suggested a potential application to muscle asymmetry assessment from an X-ray image for diagnostic and therapeutic assistance. The improved CycleGAN framework can be applied for investigating the decomposition of musculoskeletal structures from single radiographs.

A projection-domain correction method in CBCT reconstruction for head and neck radiotherapy using cycle-GAN and nonlocal means filter.

In radiation treatments for head and neck tumors, cone-beam computed tomography (CBCT) is employed for patient positioning and dose calculation of adaptive radiotherapy. However, the quality of CBCT is degraded by the scatter and noise, majorly impacting the accuracy of patient positioning and dose calculation. To improve the quality of CBCT for patients with head and neck cancer, a projection-domain CBCT correction method was proposed using a cycle-consistent generative adversarial network (cycle-GAN) and a nonlocal means filter based on a reference digitally reconstructed radiograph (DRR). A cycle-GAN was initially trained to learn mapping from CBCT projections to a DRR using the data obtained from 30 patients. For each patient, 671 CBCT projections were measured for CBCT reconstruction. Moreover, 360 Digital Reconstructed Radiographs (DRR) were computed from each patient's planning computed tomography (CT), whose projection angles ranged from 0° to 359° with an interval of 1°. By applying the trained generator of the cycle-GAN to the unseen CBCT projection, a synthetic DRR with considerably less scatter was obtained. However, annular artifacts were observed in the CBCT reconstructed with synthetic DRR. To address this issue, a nonlocal means filter based on reference DRR was used to further correct the synthetic DRR, which corrected the synthetic DRR using the calculated DRR as a reference image. Finally, the CBCT with no annular artifact and little noise was reconstructed with the corrected synthetic DRR. The proposed method was tested using the data of six patients. The corrected synthetic DRR and CBCT were compared with the corresponding real DRR and CT images. The structural preservation ability of the proposed method was evaluated using the Dice coefficients of the automatically extracted nasal cavity. Moreover, The image quality of CBCT corrected with the proposed method was objectively assessed with an five-point human scoring system and compared with CT, original CBCT and CBCT corrected with other strategies. The mean absolute value (MAE) of the relative error between the corrected synthetic and real DRR was <8%. The MAE between the corrected CBCT and corresponding CT was <30 HU. Moreover, the Dice coefficient of nasal cavity between the corrected CBCT image and the original image exceeded 98.8 for all the patients. Last but not least, the objective assessment of image qualityshowed the proposed method had an average score of 4.2 in overall image quality, which was higher than that of the original CBCT, CBCT reconstructed with synthetic DRR, and CBCT reconstructed with projections filtered with NLMF only. The proposed method can considerably improve the CBCT image quality with little anatomical distortion, improving the accuracy of radiotherapy for head and neck patients. This article is protected by copyright. All rights reserved.

Combining physics-based models with deep learning image synthesis and uncertainty in intraoperative cone-beam CT of the brain.

Image-guided neurosurgery requires high localization and registration accuracy to enable effective treatment and avoid complications. However, accurate neuronavigation based on preoperative magnetic resonance (MR) or computed tomography (CT) images is challenged by brain deformation occurring during the surgical intervention. To facilitate intraoperative visualization of brain tissues and deformable registration with preoperative images, a 3D deep learning (DL) reconstruction framework (termed DL-Recon) was proposed for improved intraoperative cone-beam CT (CBCT) image quality. The DL-Recon framework combines physics-based models with deep learning CT synthesis and leverages uncertainty information to promote robustness to unseen features. A 3D generative adversarial network (GAN) with a conditional loss function modulated by aleatoric uncertainty was developed for CBCT-to-CT synthesis. Epistemic uncertainty of the synthesis model was estimated via Monte Carlo (MC) dropout. Using spatially varying weights derived from epistemic uncertainty, the DL-Recon image combines the synthetic CT with an artifact-corrected filtered back-projection (FBP) reconstruction. In regions of high epistemic uncertainty, DL-Recon includes greater contribution from the FBP image. Twenty paired real CT and simulated CBCT images of the head were used for network training and validation, and experiments evaluated the performance of DL-Recon on CBCT images containing simulated and real brain lesions not present in the training data. Performance among learning- and physics-based methods was quantified in terms of structural similarity (SSIM) of the resulting image to diagnostic CT and Dice similarity metric (DSC) in lesion segmentation compared to ground truth. A pilot study was conducted involving seven subjects with CBCT images acquired during neurosurgery to assess the feasibility of DL-Recon in clinical data. CBCT images reconstructed via FBP with physics-based corrections exhibited the usual challenges to soft-tissue contrast resolution due to image non-uniformity, noise, and residual artifacts. GAN synthesis improved image uniformity and soft-tissue visibility but was subject to error in the shape and contrast of simulated lesions that were unseen in training. Incorporation of aleatoric uncertainty in synthesis loss improved estimation of epistemic uncertainty, with variable brain structures and unseen lesions exhibiting higher epistemic uncertainty. The DL-Recon approach mitigated synthesis errors while maintaining improvement in image quality, yielding 15%-22% increase in SSIM (image appearance compared to diagnostic CT) and up to 25% increase in DSC in lesion segmentation compared to FBP. Clear gains in visual image quality were also observed in real brain lesions and in clinical CBCT images. DL-Recon leveraged uncertainty estimation to combine the strengths of DL and physics-based reconstruction and demonstrated substantial improvements in the accuracy and quality of intraoperative CBCT. The improved soft-tissue contrast resolution could facilitate visualization of brain structures and support deformable registration with preoperative images, further extending the utility of intraoperative CBCT in image-guided neurosurgery.

Compensation cycle consistent generative adversarial networks (Comp-GAN) for synthetic CT generation from MR scans with truncated anatomy.

MR scans used in radiotherapy can be partially truncated due to the limited field of view (FOV), affecting dose calculation accuracy in MR-based radiation treatment planning. We proposed a novel Compensation-cycleGAN (Comp-cycleGAN) by modifying the cycle-consistent generative adversarial network (cycleGAN), to simultaneously create synthetic CT (sCT) images and compensate the missing anatomy from the truncated MR images. Computed tomography (CT) and T1 MR images with complete anatomy of 79 head-and-neck patients were used for this study. The original MR images were manually cropped 10-25mm off at the posterior head to simulate clinically truncated MR images. Fifteen patients were randomly chosen for testing and the rest of the patients were used for model training and validation. Both the truncated and original MR images were used in the Comp-cycleGAN training stage, which enables the model to compensate for the missing anatomy by learning the relationship between the truncation and known structures. After the model was trained, sCT images with complete anatomy can be generated by feeding only the truncated MR images into the model. In addition, the external body contours acquired from the CT images with full anatomy could be an optional input for the proposed method to leverage the additional information of the actual body shape for each test patient. The mean absolute error (MAE) of Hounsfield units (HU), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM) were calculated between sCT and real CT images to quantify the overall sCT performance. To further evaluate the shape accuracy, we generated the external body contours for sCT and original MR images with full anatomy. The Dice similarity coefficient (DSC) and mean surface distance (MSD) were calculated between the body contours of sCT and original MR images for the truncation region to assess the anatomy compensation accuracy. The average MAE, PSNR, and SSIM calculated over test patients were 93.1 HU/91.3 HU, 26.5dB/27.4dB, and 0.94/0.94 for the proposed Comp-cycleGAN models trained without/with body-contour information, respectively. These results were comparable with those obtained from the cycleGAN model which is trained and tested on full-anatomy MR images, indicating the high quality of the sCT generated from truncated MR images by the proposed method. Within the truncated region, the mean DSC and MSD were 0.85/0.89 and 1.3/0.7mm for the proposed Comp-cycleGAN models trained without/with body contour information, demonstrating good performance in compensating the truncated anatomy. We developed a novel Comp-cycleGAN model that can effectively create sCT with complete anatomy compensation from truncated MR images, which could potentially benefit the MRI-based treatment planning.

A Deep Learning Approach for Detecting Stroke from Brain CT Images Using OzNet.

A brain stroke is a life-threatening medical disorder caused by the inadequate blood supply to the brain. After the stroke, the damaged area of the brain will not operate normally. As a result, early detection is crucial for more effective therapy. Computed tomography (CT) images supply a rapid diagnosis of brain stroke. However, while doctors are analyzing each brain CT image, time is running fast. This circumstance may lead to result in a delay in treatment and making errors. Therefore, we targeted the utilization of an efficient artificial intelligence algorithm in stroke detection. In this paper, we designed hybrid algorithms that include a new convolution neural networks (CNN) architecture called OzNet and various machine learning algorithms for binary classification of real brain stroke CT images. When we classified the dataset with OzNet, we acquired successful performance. However, for this target, we combined it with a minimum Redundancy Maximum Relevance (mRMR) method and Decision Tree (DT), k-Nearest Neighbors (kNN), Linear Discriminant Analysis (LDA), Naïve Bayes (NB), and Support Vector Machines (SVM). In addition, 4096 significant features were obtained from the fully connected layer of OzNet, and we reduced the dimension of features from 4096 to 250 using the mRMR method. Finally, we utilized these machine learning algorithms to classify important features. As a result, OzNet-mRMR-NB was an excellent hybrid algorithm and achieved an accuracy of 98.42% and AUC of 0.99 to detect stroke from brain CT images.

Task-based validation and application of a scanner-specific CT simulator using an anthropomorphic phantom.

Quantitative analysis of computed tomography (CT) images traditionally utilizes real patient data that can pose challenges with replicability, efficiency, and radiation exposure. Instead, virtual imaging trials (VITs) can overcome these hurdles through computer simulations of models of patients and imaging systems. DukeSim is a scanner-specific CT imaging simulator that has previously been validated with simple cylindrical phantoms, but not with anthropomorphic conditions and clinically relevant measurements. To validate a scanner-specific CT simulator (DukeSim) for the assessment of lung imaging biomarkers under clinically relevant conditions across multiple scanners using an anthropomorphic chest phantom, and to demonstrate the utility of virtual trials by studying the effects or radiation dose and reconstruction kernels on the lung imaging quantifications. An anthropomorphic chest phantom with customized tube inserts was imaged with two commercial scanners (Siemens Force and Siemens Flash) at 28 dose and reconstruction conditions. A computational version of the chest phantom was used with a scanner-specific CT simulator (DukeSim) to simulate virtual images corresponding to the settings of the real acquisitions. Lung imaging biomarkers were computed from both real and simulated CT images and quantitatively compared across all imaging conditions. The VIT framework was further utilized to investigate the effects of radiation dose (20-300mAs) and reconstruction settings (Qr32f, Qr40f, and Qr69f reconstruction kernels using ADMIRE strength 3) on the accuracy of lung imaging biomarkers, compared against the ground-truth values modeled in the computational chest phantom. The simulated CT images matched closely the real images for both scanners and all imaging conditions qualitatively and quantitatively, with the average biomarker percent error of 3.51% (range 0.002%-18.91%). The VIT study showed that sharper reconstruction kernels had lower accuracy with errors in mean lung HU of 84-94HU, lung volume of 797-3785cm3 , and lung mass of -800 to 1751g. Lower tube currents had the lower accuracy with errors in mean lung HU of 6-84HU, lung volume of 66-3785cm3 , and lung mass of 170-1751g. Other imaging biomarkers were consistent under the studied reconstruction settings and tube currents. We comprehensively evaluated the realism of DukeSim in an anthropomorphic setup across a diverse range of imaging conditions. This study paves the way toward utilizing VITs more reliably for conducting medical imaging experiments that are not practical using actual patient images.

Multimodal image translation via deep learning inference model trained in video domain

BackgroundCurrent medical image translation is implemented in the image domain. Considering the medical image acquisition is essentially a temporally continuous process, we attempt to develop a novel image translation framework via deep learning trained in video domain for generating synthesized computed tomography (CT) images from cone-beam computed tomography (CBCT) images.MethodsFor a proof-of-concept demonstration, CBCT and CT images from 100 patients were collected to demonstrate the feasibility and reliability of the proposed framework. The CBCT and CT images were further registered as paired samples and used as the input data for the supervised model training. A vid2vid framework based on the conditional GAN network, with carefully-designed generators, discriminators and a new spatio-temporal learning objective, was applied to realize the CBCT–CT image translation in the video domain. Four evaluation metrics, including mean absolute error (MAE), peak signal-to-noise ratio (PSNR), normalized cross-correlation (NCC), and structural similarity (SSIM), were calculated on all the real and synthetic CT images from 10 new testing patients to illustrate the model performance.ResultsThe average values for four evaluation metrics, including MAE, PSNR, NCC, and SSIM, are 23.27 ± 5.53, 32.67 ± 1.98, 0.99 ± 0.0059, and 0.97 ± 0.028, respectively. Most of the pixel-wise hounsfield units value differences between real and synthetic CT images are within 50. The synthetic CT images have great agreement with the real CT images and the image quality is improved with lower noise and artifacts compared with CBCT images.ConclusionsWe developed a deep-learning-based approach to perform the medical image translation problem in the video domain. Although the feasibility and reliability of the proposed framework were demonstrated by CBCT–CT image translation, it can be easily extended to other types of medical images. The current results illustrate that it is a very promising method that may pave a new path for medical image translation research.

An unsupervised domain adaptation brain CT segmentation method across image modalities and diseases

Computed tomography (CT) is the primary diagnostic tool for brain diseases. To determine the appropriate treatment plan, it is necessary to ascertain the patient’s bleeding volume. Automatic segmentation algorithms for hemorrhagic lesions can significantly improve efficiency and avoid treatment delays. However, for deep supervised learning algorithms, a large amount of labeled training data is usually required, making them difficult to apply clinically. In this study, we propose an unsupervised domain adaptation method that is an unsupervised domain adaptation segmentation model that can be trained across modalities and diseases. We call it AMD-DAS for brain CT hemorrhage segmentation tasks. This circumvents the heavy data labeling task by converting the source domain data (MRI with glioma) to our task’s required data (CT with Intraparenchymal hemorrhage (IPH)). Our model implements a two-stage domain adaptation process to achieve this objective. In the first stage, we train a pseudo-CT image synthesis network using the CycleGAN architecture through a matching mechanism and domain adaptation approach. In the second stage, we use the model trained in the first stage to synthesize the pseudo-CT images. We use the pseudo-CT with source domain labels and real CT images to train a domain-adaptation segmentation model. Our method exhibits a better performance than the basic one-stage domain adaptation segmentation method (+11.55 Dice score) and achieves an 86.93 Dice score in the IPH unsupervised segmentation task. Our model can be trained without using a ground-truth label, therefore increasing its application potential. Our implementation is publicly available at https://github.com/GuanghuiFU/AMD-DAS-Brain-CT-Segmentation.

Feasibility analysis on simultaneous electron density and attenuation coefficient reconstruction.

Creating a viable reconstruction method for Compton scatter tomography remains challenging. Accounting for scatter attenuation when the underlying attenuation map is not known is particularly challenging, and current mathematical approaches to this vary widely. This work explores a novel approach to joint scatter and attenuation image reconstruction, which leverages the underlying structural similarity between the two images and incorporates a deep learning model in an alternating iterative reconstruction scheme. A single-view computed tomography (CT) imaging procedure for recording Compton scatter is first described. A joint reconstruction model, which iterates between algebraically reconstructing scatter images and estimating the attenuation via deep learning, is then proposed. This model is tested on both a generated dataset of 2D phantom images designed to mimic human tissues as well as a realistically simulated dataset based on real CT images. Testing results yield convergence of the model and decent reconstruction quality to distinguish crucial features such as tumors and lesions, demonstrating the potential principled utilities of this configuration and deep learning approach. The model achieved a structural similarity index measure of at least 0.82 for scatter and 0.88 for attenuation reconstructions with the realistically simulated dataset. The iterative, deep learning approach outlined in this work shows potential for future efficient medical imaging procedures, reconstructing images with limited scatter information.

Unsupervised domain adaptation based COVID-19 CT infection segmentation network.

Automatic segmentation of infection areas in computed tomography (CT) images has proven to be an effective diagnostic approach for COVID-19. However, due to the limited number of pixel-level annotated medical images, accurate segmentation remains a major challenge. In this paper, we propose an unsupervised domain adaptation based segmentation network to improve the segmentation performance of the infection areas in COVID-19 CT images. In particular, we propose to utilize the synthetic data and limited unlabeled real COVID-19 CT images to jointly train the segmentation network. Furthermore, we develop a novel domain adaptation module, which is used to align the two domains and effectively improve the segmentation network’s generalization capability to the real domain. Besides, we propose an unsupervised adversarial training scheme, which encourages the segmentation network to learn the domain-invariant feature, so that the robust feature can be used for segmentation. Experimental results demonstrate that our method can achieve state-of-the-art segmentation performance on COVID-19 CT images.

Realistic High-Resolution Body Computed Tomography Image Synthesis by Using Progressive Growing Generative Adversarial Network: Visual Turing Test.

BackgroundGenerative adversarial network (GAN)–based synthetic images can be viable solutions to current supervised deep learning challenges. However, generating highly realistic images is a prerequisite for these approaches.ObjectiveThe aim of this study was to investigate and validate the unsupervised synthesis of highly realistic body computed tomography (CT) images by using a progressive growing GAN (PGGAN) trained to learn the probability distribution of normal data.MethodsWe trained the PGGAN by using 11,755 body CT scans. Ten radiologists (4 radiologists with <5 years of experience [Group I], 4 radiologists with 5-10 years of experience [Group II], and 2 radiologists with >10 years of experience [Group III]) evaluated the results in a binary approach by using an independent validation set of 300 images (150 real and 150 synthetic) to judge the authenticity of each image.ResultsThe mean accuracy of the 10 readers in the entire image set was higher than random guessing (1781/3000, 59.4% vs 1500/3000, 50.0%, respectively; P<.001). However, in terms of identifying synthetic images as fake, there was no significant difference in the specificity between the visual Turing test and random guessing (779/1500, 51.9% vs 750/1500, 50.0%, respectively; P=.29). The accuracy between the 3 reader groups with different experience levels was not significantly different (Group I, 696/1200, 58.0%; Group II, 726/1200, 60.5%; and Group III, 359/600, 59.8%; P=.36). Interreader agreements were poor (κ=0.11) for the entire image set. In subgroup analysis, the discrepancies between real and synthetic CT images occurred mainly in the thoracoabdominal junction and in the anatomical details.ConclusionsThe GAN can synthesize highly realistic high-resolution body CT images that are indistinguishable from real images; however, it has limitations in generating body images of the thoracoabdominal junction and lacks accuracy in the anatomical details.

Dynamically learned PSO based neighborhood influenced fuzzy c-means for pre-treatment and post-treatment organ segmentation from CT images

Background and ObjectiveThe accurate segmentation of pre-treatment and post-treatment organs is always perceived as a challenging task in medical image analysis field. Especially, in those situations where the amount of data set is limited, the researchers are compelled to design unsupervised model for segmentation. In this paper, we propose a novel dynamically learned particle swarm optimization based neighborhood influenced fuzzy c-means (DLPSO-NIFCM) clustering (unsupervised learning model) for solving pre-treatment and post-treatment organs segmentation problems. The proposed segmentation technique has been successfully applied to segment the liver parts from the Computed Tomography (CT) images of abdomen and also the lung parenchyma from the lungs CT images. MethodologyIn the proposed method, we formulate a primary convex objective function by considering the membership value of a pixel as well as the membership of its other neighboring pixels. Then we apply a new algebraic transformation on the primary objective function to design a new and more suitable objective function without losing convexity of the primary objective function. This new objective function is compatible for hybridization with any heuristic search technique in true sense. In this work, we propose a dynamically learned PSO to obtain the initial cluster centroids from the final objective function. Finally, we use a graph-based isolation mechanism for refining the segmentation results. Results and ConclusionThis hybrid method, along with the restructured single variable objective function of the distance, leads to accurate clustering results with relatively lesser converging time as compared to the state-of-the-art methods. The segmentation results, obtained through several experiments with real CT images, are encouraging. The numerical values of different performance metrics obtained over the same data set confirm that the proposed algorithm performs better with respect to the state-of-the-art methods. Hence, we may consider the proposed method as a promising tool for clustering and CT image segmentation in a Computer Aided Diagnostic (CAD) system.

Algorithm for post-processing of tomography images to calculate the dimension-geometric features of porous structures

An algorithm for post-processing of the grayscale 3D computed tomography (CT) images of porous structures with the automatic selection of filtering parameters is proposed. The determination of parameters is carried out on a representative part of the image under analysis. A criterion for the search for optimal filtering parameters based on the count of "levitating stone" voxels is described. The stages of CT image filtering and its binarization are performed sequentially. Bilateral and anisotropic diffuse filtering is implemented; the Otsu method for unbalanced classes is chosen for binarization. Verification of the proposed algorithm was carried out on model data. To create model porous structures, we used our image generator, which implements the function of anisotropic porous structures generation. Results of the post-processing of real CT images containing noise and reconstruction artifacts by the proposed method are discussed.

Deep learning for COVID-19 chest CT (computed tomography) image analysis: A lesson from lung cancer.

As a recent global health emergency, the quick and reliable diagnosis of COVID-19 is urgently needed. Thus, many artificial intelligence (AI)-base methods are proposed for COVID-19 chest CT (computed tomography) image analysis. However, there are very limited COVID-19 chest CT images publicly available to evaluate those deep neural networks. On the other hand, a huge amount of CT images from lung cancer are publicly available. To build a reliable deep learning model trained and tested with a larger scale dataset, the proposed model builds a public COVID-19 CT dataset, containing 1186 CT images synthesized from lung cancer CT images using CycleGAN. Additionally, various deep learning models are tested with synthesized or real chest CT images for COVID-19 and Non-COVID-19 classification. In comparison, all models achieve excellent results in accuracy, precision, recall and F1 score for both synthesized and real COVID-19 CT images, demonstrating the reliable of the synthesized dataset. The public dataset and deep learning models can facilitate the development of accurate and efficient diagnostic testing for COVID-19.

Time-distanced gates in long short-term memory networks.

The Long Short-Term Memory (LSTM) network is widely used in modeling sequential observations in fields ranging from natural language processing to medical imaging. The LSTM has shown promise for interpreting computed tomography (CT) in lung screening protocols. Yet, traditional image-based LSTM models ignore interval differences, while recently proposed interval-modeled LSTM variants are limited in their ability to interpret temporal proximity. Meanwhile, clinical imaging acquisition may be irregularly sampled, and such sampling patterns may be commingled with clinical usages. In this paper, we propose the Distanced LSTM (DLSTM) by introducing time-distanced (i.e., time distance to the last scan) gates with a temporal emphasis model (TEM) targeting at lung cancer diagnosis (i.e., evaluating the malignancy of pulmonary nodules). Briefly, (1) the time distance of every scan to the last scan is modeled explicitly, (2) time-distanced input and forget gates in DLSTM are introduced across regular and irregular sampling sequences, and (3) the newer scan in serial data is emphasized by the TEM. The DLSTM algorithm is evaluated with both simulated data and real CT images (from 1794 National Lung Screening Trial (NLST) patients with longitudinal scans and 1420 clinical studied patients). Experimental results on simulated data indicate the DLSTM can capture families of temporal relationships that cannot be detected with traditional LSTM. Cross-validation on empirical CT datasets demonstrates that DLSTM achieves leading performance on both regularly and irregularly sampled data (e.g., improving LSTM from 0.6785 to 0.7085 on F1 score in NLST). In external-validation on irregularly acquired data, the benchmarks achieved 0.8350 (CNN feature) and 0.8380 (with LSTM) on AUC score, while the proposed DLSTM achieves 0.8905. In conclusion, the DLSTM approach is shown to be compatible with families of linear, quadratic, exponential, and log-exponential temporal models. The DLSTM can be readily extended with other temporal dependence interactions while hardly increasing overall model complexity.

Scaling properties of binary and greyscale images in the context of X-ray soil tomography

Characterization of the complex soil structure is one the cornerstones of soil science and pore space detection is a crucial step in this process. Synthetic soil image construction has been proved to be an efficient resource for validating different binarization methods given that, unlike in real world, ground truth information is known. In this work, we introduce an improved Truncated Multifractal Method (TMM), to better simulate synthetic computed tomography (CT) soil images and then we generate 150 synthetic images with three different porosities (7%, 12% and 17%), both in greyscale and in binary scale (pore spaces). Synthetic images are then compared with two sets of 260 slides of real CT soil images, in order to validate the goodness of the method. All images are subjected to multifractal analysis where we show a detailed comparative analysis of parameters such as lacunarity, characteristic length and multifractal spectrum, that are calculated both for the whole set of synthetic (greyscale and binary) and for the sets of real CT soil images. With respect to lacunarity, a not previously reported inverse relationship between binary and grey lacunarity is found for this range of porosities. Moreover, we have also reported a new relationship between lacunarity and characteristic length. Similar multifractal results, that we obtain for real CT and synthetic soil images, prove that TMM is a reasonable solution to create simulated CT soil images. Finally, a segmentation test was carried out, using TMM synthetic greyscale soil images and its binary counterpart as ground-truth information, evaluating global (Otsu) and local (Combining Singularity-CA) binarization methods, where we report better performance for the last.