Coronary Artery Calcium Scoring (CACS) is used for cardiac risk assessment caused by atherosclerotic plaque or other coronary artery diseases. Images from Non-Contrast (NC) cardiac Computed Tomography (CT) scans acquired at 120kVp are used in computing Agatston scoring for CACS. These scans, if done at lower peak voltage can reduce X-Ray radiation exposure. This, however, changes CT attenuation values for all tissues, as well as calcification compared to 120kVp scan, thus making it unusable for Agatston scoring. We propose a learning-based method to translate a CT image acquired at lower kVp to a 120kVp equivalent image, such that the same calcium scoring protocol can be used on these scans. We establish that the proposed method enables appropriate translation of CT values in calcification regions, thereby allowing similar calcium score (error < 6%) for a patient at reduced dose. Our proposed learning-based approach shows robust performance across datasets.
Following the acquisition of images in CT, a crucial post-processing step involves orienting the volumetric image to align with standard viewing planes, facilitating the assessment of disease extent and other pathologies. However, manual alignment is not only time-consuming but can also pose challenges in achieving consistent standard plane views, particularly for lesser skilled technologists. Existing automated solutions, primarily based on registration techniques, encounter reduced accuracy in cases involving significant rotations, pediatric patients, and instances with pronounced pathological effects. This limitation arises due to their reliance on symmetry. In severe scenarios, registration-based methods can exacerbate image misalignment compared to the original input. To address these concerns, this study introduces a landmark-based automated image alignment method. This method presents three key advantages: robust alignment across diverse data variations, the capability to identify algorithm failures and gracefully terminate, and the ability to align images with different standard planes. The effectiveness of our method is showcased through a comparative evaluation with registration-based approaches. The evaluation employs a test dataset comprising various head cases across different age groups, reaffirming the effectiveness of our proposed method.
Kernel Synthesis (KS) in CT is an advanced image processing technique that involves the conversion of an image acquired with one specific convolution kernel into an image that appears as if it were acquired with a different kernel. This process doesn't require raw sinogram or reconstruction algorithm to generate images. While kernel synthesis methods exhibit promise, they can introduce noise and artifacts during transformation, highlighting the importance of effective noise and artifact management in the input data before synthesis. In this work, we represent kernel synthesis as a deep neural network regularized inverse problem. We optimize the convolutional neural network (CNN) weights in a data-consistent manner where CNN acts as an implicit prior to regularize the solution. Since, the learned CNN priors are more generic than hand crafted priors (like sparsity and total-variation), they helps control artifacts while preserving the details in the image. Experimental results for a low-resolution kernel (STND) data to a high resolution kernel (LUNG) data conversion clearly indicates that the output images from the proposed method resemble closely with images reconstructed using high-resolution kernel (LUNG) while simultaneously reducing the noise and clutter.
Pinwheel artifact is caused in multidetector Computed Tomography (CT) helical scans due to under-sampling of z-planes during thin slice reconstruction. There are a few hardware-based solutions like flying focal spot (FFS), which aims at generating more samples during the acquisition, thus enabling aliasing free thin slice images. However, these methods are expensive both in manufacturing capability, as well as impact on hardware life. Deep learning (DL) based methods have shown significant improvement in pin-wheel artifact reduction. Most DL-based methods use images generated from FFS or similar hardware-based enhancement to train the deep learning network, thus restricting usability of these methods on systems without these hardware enhancements. This work proposes a novel DL method to generate pin-wheel free thin-slice images from helical scans for systems not equipped with these hardware capabilities. Artifact-free thin slice images, which are used as targets for artifact reduction network are generated through DL-based super-resolution along z-direction from thick slice images reconstructed from the same scan. The framework is trained with ~16000 coronal/sagittal slices from GE-Revolution system. Clinical image review and statistical analysis of the inferencing results have shown significant artifact reduction and improved diagnostic image quality while reducing noise. A Likert score study shows significant enhancement of proposed method over other image processing-based solutions available.
Low performing pixels (LPP)/missing/bad channels in CT detectors, if left uncorrected cause ring and streak artifacts, structured non-uniformities, and make the reconstructed image unusable for diagnostic purposes. Many image processing methods are proposed to correct the ring and streak artifacts in reconstructed images, but it is more appropriate to correct the LPPs in sinogram domain as the errors are localized. Although Generative Adversarial Networks based sinogram inpainting methods have shown promise in interpolating the missing sinogram information, it is often observed that the reconstructed images lack diagnostic value especially in visualizing soft tissues with certain window width and level. In this work, we propose a deep-learning based solution that operates on the sinogram data to remove the distortions cause by LPPs. This method leverages the CT system geometry (including conjugate ray information) to learn the anatomy aware interpolation in the sinogram domain. We demonstrated the efficacy of the proposed method using data acquired on GE RevACT multi-slice CT system with flat-panel detector. We have considered 46 axial head scans out of them 42 sets are used for training and the remaining 4 sets for validation/testing. We have simulated isolated LPPs accounting for 10% of total channels in the central panel of the detector and corrected them using the proposed approach. Detailed statistical analysis has revealed that, approximately 5dB improvement in SNR is observed in both sinogram and reconstruction domain as compared to classical bicubic and Lagrange interpolation methods. Also, with reduction in ring and streak artifacts, the perceptual image quality is improved across all the test images.
In Computed Tomography (CT), problems like low to high dose image conversion, denoising and super-resolution using deep regression network have gained a lot of importance. This led to the study of multiple denoising approaches with these networks, as non-linear characteristics of the network often changes the noise pattern. Noise2Noise (N2N) and similar studies have shown the impact of these networks, where one can use noisy image pairs and exploit the uncorrelated nature of noise to generate denoised images. In this paper, we study the behaviour of regression network for a domain translation between one energy image to another energy and its impact on noise. Inter-kVp translation leads to change in attenuation values in CT image. A design of experiments is set up using CATSim phantom with different tissue types at varying levels of density and proportions of materials, for heart, soft tissue, fat, calcification and bone. The intent is to understand the impact of inter and intra-tissue dynamic ranges, as this network learns intensity translation of the image between 2 kVp levels and noise characteristics. The results demonstrate ability of regression network to change intensty values from low kVp image to high kVp image. It also shows impact on noise level in a tissue is proportional to (a) intra-tissue variability and (b) the desired mean shift of the tissue between two energy levels. Simultaneously, there is a marked change in the artifacts and resultant image quality, as expected, through the learning method.
CT systems with large detector size suffer from lower z-resolution leading to pixelated images and inability to detect small structures thus adversely impacting the diagnosis and screening. Overlap reconstruction can partially reduce the stair-step artifacts but does not improve the effect of wider slice sensitivity profile (SSP) and thus continues to have reduced visibility of smaller structures. In this work, we propose a supervised deep learning method for z-resolution enhancement such that (a) the effective SSP of resulting image is reduced, (b) quantitative values of tissue (CT numbers) and tissue-contrast are preserved; (c) very limited noise enhancement and (d) improved tissue interface in bone/soft tissue. The proposed method devises a super resolution (SURE) network which is trained to map the low resolution (LR) slices to the corresponding high resolution (HR) slices. A 2D network is trained with sagittal and coronal slices with the LR-HR pair sets. The training is performed using ground truth HR slices obtained from high end systems, and the corresponding LR slices are synthesized by either using retro reconstruction with higher slice thickness and spacing or through averaging of slices in z-direction from HR images. The network is trained with both these types of images with helical acquisition volumes from a range of scanners. Qualitative and quantitative analysis is done on the predicted HR images and compared with the original HR images. FWHM for SSP of the predicted HR images reduced from ~0.98 to ~0.73, when the target was 0.64, thus improving the real z-resolution. HU distribution of different tissue types also showed stability in terms of mean value. Noise measured through standard deviation was slightly higher than the LR image but lower than that of original HR images. PSNR also showed consistent improvement on all the cases across 3 different systems.
Images produced by CT systems with larger detector pixels often suffer from lower z resolution due to their wider slice sensitivity profile (SSP). Reducing the effect of SSP blur will result in resolution of finer structures and enables better clinical diagnosis. Hardware solutions such as dicing the detector cells smaller or dynami- cally deflecting the X-ray focal spot do exist to improve the resolution, but they are expensive. In the past, algorithmic solutions like deconvolution techniques also have been used to reduce the SSP blur. These model- based approaches are iterative in nature and are time consuming. Recently, supervised data-driven deep learning methods have become popular in computer vision for deblurring/deconvolution applications. Though most of these methods need corresponding pairs of blurred (LR) and sharp (HR) images, they are non-iterative during inference and hence are computationally efficient. However, unlike the model-based approaches, these methods do not explicitly model the physics of degradation. In this work, we propose Resolution Amelioration using Machine Adaptive Network (RAMAN), a self-supervised deep learning framework, that explicitly uses best of both learning and model based approaches. The framework explicitly accounts for the physics of degradation and appropriately regularizes the learning process. Also, in contrary to supervised deblurring methods that need paired LR and HR images, the RAMAN framework requires only LR images and SSP information for training, making it self-supervised. Validation of proposed framework with images obtained from larger detector systems shows marked improvement in image sharpness while maintaining HU integrity.
Artificial intelligence (AI) models are used in medical image processing and analysis tasks like organ segmentation, anomaly detection, image reconstruction, and so on. Most often these models are trained on specific type of source domain images (non-contrast or contrast enhanced, specific field-of-view (FOV), dosage, demography, etc). It is desirable to adapt these models to a different but similar target domain, through unsupervised or semi-supervised learning methods. This paper describes a framework to re-purpose trained organ segmentation models for a target domain on which the model was not trained. The adaptation is proposed using an additional autoencoder network as a post-processing step to improve accuracy of predicted segmentation mask. Unsupervised and semi-supervised versions of adaptation are tested on contrast enhanced Computed Tomography (CT) liver, cardiac and lung exams. Experiment results show adaptability of a trained network from non-contrast to contrast enhanced scans with improved accuracy in contrast enhanced volume segmentation. A domain adaptation case study on lung disease exams with bacterial pneumonia or COVID-19 pathology also shows the effectiveness of proposed methodology.
Automated segmentation of vertebral bone from diagnostic Computed Tomography (CT) images has become an important part of clinical workflow today. There is an increasing need for computer aided diagnosis applications of various spine disorders including scoliosis, fracture detection and even automated reporting. While modelbased methods have been widely used, recent deep Learning methods have shown a great potential in this area. However, choice of optimal configuration of the network to get the best segmentation performance is challenging. In this work, we explore the impact of different training and inference options, including dimensions, activation function, batch normalization, kernel size, filters, patch size and patch selection strategy in U-Net architecture. 20 publicly available CT Spine datasets from Spineweb repository was used in this study divided into training/test datasets. Training with different DL configurations were repeated with these datasets. We used the best weights corresponding to each configuration for inference on the independent test dataset. These results on the test dataset with the best weights for each configurations were compared. 3D models performed consistently better than 2D approaches. Overlapped patch based inference had a big impact on enhancing performance accuracy. The selection of training patch size was also found to be crucial in improving the model performance. Moreover, the need for an effective balance of positive and negative training patches was found. The best performance in our study was obtained by using overlapped patch inference, training with RELU activation and batch normalization in a 3D U-Net architecture with training patch size of 128×128×32 that resulted in average values of precision= 97%, sensitivity= 96% and F1 (Dice)= 96% for the test dataset.
Parsing volumetric computed tomography (CT) into 10 or more salient organs simultaneously is a challenging task with many applications such as personalized scan planning and dose reporting. In the clinic, pre-scan data can come in the form of very low dose volumes acquired just prior to the primary scan or from an existing primary scan. To localize organs in such diverse data, we propose a new learning based framework that we call hierarchical pictorial structures (HPS) which builds multiple levels of models in a tree-like hierarchy that mirrors the natural decomposition of human anatomy from gross structures to finer structures. Each node of our hierarchical model learns (1) the local appearance and shape of structures, and (2) a generative global model that learns probabilistic, structural arrangement. Our main contribution is twofold. First we embed the pictorial structures approach in a hierarchical framework which reduces test time image interpretation and allows for the incorporation of additional geometric constraints that robustly guide model fitting in the presence of noise. Second we guide our HPS framework with the probabilistic cost maps extracted using random decision forests using volumetric 3D HOG features which makes our model fast to train and fast to apply to novel test data and posses a high degree of invariance to shape distortion and imaging artifacts. All steps require approximate 3 mins to compute and all organs are located with suitably high accuracy for our clinical applications such as personalized scan planning for radiation dose reduction. We assess our method using a database of volumetric CT scans from 81 subjects with widely varying age and pathology and with simulated ultra-low dose cadaver pre-scan data.
Tissue characterization from imaging studies is an integral part of clinical practice. We describe a spectral filter
design for tissue separation in dual energy CT scans obtained from Gemstone Spectral Imaging scanner. It
enables to have better 2D/3D visualization and tissue characterization in normal and pathological conditions.
The major challenge to classify tissues in conventional computed tomography (CT) is the x-ray attenuation
proximity of multiple tissues at any given energy. The proposed method analyzes the monochromatic images
at different energy levels, which are derived from the two scans obtained at low and high KVp through fast
switching. Although materials have a distinct attenuation profile across different energies, tissue separation
is not trivial as tissues are a mixture of different materials with range of densities that vary across subjects.
To address this problem, we define spectral filtering, that generates probability maps for each tissue in multi-energy
space. The filter design incorporates variations in the tissue due to composition, density of individual
constituents and their mixing proportions. In addition, it also provides a framework to incorporate zero mean
Gaussian noise. We demonstrate the application of spectral filtering for bone-free vascular visualization and
calcification characterization.
Pulmonary fissures separate human lungs into five distinct regions called lobes. Detection of fissure is essential
for localization of the lobar distribution of lung diseases, surgical planning and follow-up. Treatment planning
also requires calculation of the lobe volume. This volume estimation mandates accurate segmentation of the
fissures. Presence of other structures (like vessels) near the fissure, along with its high variational probability in
terms of position, shape etc. makes the lobe segmentation a challenging task. Also, false incomplete fissures and
occurrence of diseases add to the complications of fissure detection. In this paper, we propose a semi-automated
fissure segmentation algorithm using a minimal path approach on CT images. An energy function is defined such
that the path integral over the fissure is the global minimum. Based on a few user defined points on a single slice
of the CT image, the proposed algorithm minimizes a 2D energy function on the sagital slice computed using (a)
intensity (b) distance of the vasculature, (c) curvature in 2D, (d) continuity in 3D. The fissure is the infimum
energy path between a representative point on the fissure and nearest lung boundary point in this energy domain.
The algorithm has been tested on 10 CT volume datasets acquired from GE scanners at multiple clinical sites.
The datasets span through different pathological conditions and varying imaging artifacts.
KEYWORDS: Image registration, Image segmentation, Medical imaging, Image processing, Current controlled current source, Computed tomography, Image processing algorithms and systems, Detection and tracking algorithms
Automated labeling of the bronchial tree is essential for localization of airway related diseases (e.g. chronic bronchitis) and is also a useful precursor to lung-lobe labeling. We describe an automated method for registration-based labeling of a bronchial tree. The bronchial tree is segmented from a CT image using a region-growing based algorithm. The medial line of the extracted tree is then computed using a potential field based approach. The expert-labeled target (atlas) and the source bronchial trees in the form of extracted centerline point sets are brought into alignment by calculating a non-rigid thin-plate spline (TPS) mapping from the source to the target. The registration takes into account global as well as local variations in anatomy between the two images through the use of separable linear and non-linear components of the transformation; as a result it is well suited to matching structures that deviate at finer levels: namely higher order branches. The method is validated by registering together pairs of datasets for which the ground truth labels are known in advance: the labels are transferred after matching target to source and then compared with the true values. The method was tested on datasets each containing 18 branch centerpoints and 12 bifurcation locations (30 landmarks in total) annotated manually by a radiologist, where the performance was measured as the number of landmarks having the correct transfer of labels. An overall accuracy of labeling of 91.5 % was obtained in matching 23 pairs of datasets obtained from different patients.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.