scholarly journals End-to-End Deep Learning CT Image Reconstruction for Metal Artifact Reduction

2021 ◽  
Vol 12 (1) ◽  
pp. 404
Author(s):  
Dominik F. Bauer ◽  
Constantin Ulrich ◽  
Tom Russ ◽  
Alena-Kathrin Golla ◽  
Lothar R. Schad ◽  
...  
Keyword(s):  
Deep Learning ◽  
Reference Method ◽  
Projection Data ◽  
Ct Reconstruction ◽  
Back Projection ◽  
Post Processing ◽  
Metal Artifacts ◽  
Reconstruction Process ◽  
Ct Guided ◽  
End To End

Metal artifacts are common in CT-guided interventions due to the presence of metallic instruments. These artifacts often obscure clinically relevant structures, which can complicate the intervention. In this work, we present a deep learning CT reconstruction called iCTU-Net for the reduction of metal artifacts. The network emulates the filtering and back projection steps of the classical filtered back projection (FBP). A U-Net is used as post-processing to refine the back projected image. The reconstruction is trained end-to-end, i.e., the inputs of the iCTU-Net are sinograms and the outputs are reconstructed images. The network does not require a predefined back projection operator or the exact X-ray beam geometry. Supervised training is performed on simulated interventional data of the abdomen. For projection data exhibiting severe artifacts, the iCTU-Net achieved reconstructions with SSIM = 0.970±0.009 and PSNR = 40.7±1.6. The best reference method, an image based post-processing network, only achieved SSIM = 0.944±0.024 and PSNR = 39.8±1.9. Since the whole reconstruction process is learned, the network was able to fully utilize the raw data, which benefited from the removal of metal artifacts. The proposed method was the only studied method that could eliminate the metal streak artifacts.

Projection-to-image transform frame: a lightweight block reconstruction network (LBRN) for computed tomography

2021 ◽  
Author(s):  
Genwei Ma ◽  
Xing Zhao ◽  
Yining Zhu ◽  
Huitao Zhang
Keyword(s):  
Neural Network ◽  
Computed Tomography ◽  
Real Data ◽  
Projection Data ◽  
Ct Reconstruction ◽  
Back Projection ◽  
Image Block ◽  
Metal Artifacts ◽  
Reconstruction Process ◽  
Memory Space

Abstract To solve the problem of learning based computed tomography (CT) reconstruction, several reconstruction networks were invented. However, applying neural network to tomographic reconstruction still remains challenging due to unacceptable memory space requirement. In this study, we presents a novel lightweight block reconstruction network (LBRN), which transforms the reconstruction operator into a deep neural network by unrolling the filter back-projection (FBP) method. Specifically, the proposed network contains two main modules, which, respectively, correspond to the filter and back-projection of FBP method. The first module of LBRN decouples the relationship of Radon transform between the reconstructed image and the projection data. Therefore, the following module, block back-projection module, can use the block reconstruction strategy. Due to each image block is only connected with part filtered projection data, the network structure is greatly simplified and the parameters of the whole network is dramatically reduced. Moreover, this approach is trained end-to-end, working directly from raw projection data and does not depend on any initial images. Five reconstruction experiments are conducted to evaluate the performance of the proposed LBRN: full angle, low-dose CT, region of interest (ROI), metal artifacts reduction and real data experiment. The results of the experiments show that the LBRN can be effectively introduced into the reconstruction process and has outstanding advantages in terms of different reconstruction problems.

scholarly journals Completion of Metal-Damaged Traces Based on Deep Learning in Sinogram Domain for Metal Artifacts Reduction in CT Images

Sensors ◽  
2021 ◽  
Vol 21 (24) ◽  
pp. 8164
Author(s):  
Linlin Zhu ◽  
Yu Han ◽  
Xiaoqi Xi ◽  
Lei Li ◽  
Bin Yan
Keyword(s):  
Computed Tomography ◽  
Deep Learning ◽  
Ct Images ◽  
Projection Data ◽  
Metal Artifacts ◽  
Deep Networks ◽  
Data Masking ◽  
Projection Images ◽  
Artifacts Reduction ◽  
The Stability

In computed tomography (CT) images, the presence of metal artifacts leads to contaminated object structures. Theoretically, eliminating metal artifacts in the sinogram domain can correct projection deviation and provide reconstructed images that are more real. Contemporary methods that use deep networks for completing metal-damaged sinogram data are limited to discontinuity at the boundaries of traces, which, however, lead to secondary artifacts. This study modifies the traditional U-net and adds two sinogram feature losses of projection images—namely, continuity and consistency of projection data at each angle, improving the accuracy of the complemented sinogram data. Masking the metal traces also ensures the stability and reliability of the unaffected data during metal artifacts reduction. The projection and reconstruction results and various evaluation metrics reveal that the proposed method can accurately repair missing data and reduce metal artifacts in reconstructed CT images.

Application of Computed Tomographic Image Reconstruction Algorithms Based on Filtered Back-Projection in Diagnosis of Bone Trauma Diseases

2020 ◽  
Vol 10 (5) ◽  
pp. 1219-1224
Author(s):  
Xianyu Li ◽  
Yulin He ◽  
Qun Hua
Keyword(s):  
Image Reconstruction ◽  
Reconstruction Algorithm ◽  
Projection Data ◽  
Ct Image ◽  
Back Projection ◽  
Post Processing ◽  
Filtered Back Projection ◽  
X Ray ◽  
Image Reconstruction Algorithm ◽  
Bone Trauma

Objective: To improve the diagnostic rate of bone trauma diseases by applying image reconstruction algorithm based on filtered back-projection to CT images of bone trauma. Methods: Sixty-three patients with bone trauma in our hospital were selected. After hospitalization, 63 patients took satisfactory localization images to make the lesions on the localization images close to or even exceed the resolution of conventional X-ray films. After scanning, the post-processing workstation software was used for post-processing of image reconstruction algorithm based on filtered back-projection. Finally, the diagnostic accuracy of X-ray plain film, common CT image and image examination based on filtered back-projection was compared statistically. Results: Among 63 cases of bone trauma, 48 cases were found by routine CT cross-sectional examination. The image reconstruction algorithm based on filtered back-projection was applied to all cases of wrist ulnar and trauma examination. The three-dimensional imaging can display the length, direction, shape of articular surface and fracture end of bone trauma as well as the size and spatial position of free small bone fragments stereoscopically and accurately. The relationship between bone trauma and placement. Discussion: Experiments show that when the projection data are complete, the filtering back-projection algorithm can reconstruct the image better, and the overall evaluation of the new filtering function is the best. Usually, the projection data are often incomplete, sometimes even seriously insufficient. At this time, it is necessary to adopt iterative reconstruction algorithm. However, no matter which algorithm is adopted, the reconstruction speed and accuracy are improved, and the quality of the reconstructed image is improved. It remains the direction of future efforts. The FBP method is the basic common algorithm for reconstructing image, and it is also the basis of many other algorithms. It is widely used in medical CT and other fields. Conclusion: The improved CT image reconstruction algorithm based on filtered back-projection has high application value in the diagnosis of bone trauma diseases. By comparing the three indexes of serial processing time, information transfer interface and image noise, the suspicious site of bone trauma can be diagnosed clearly. In recent years, with the popularization of CT and the emergence of spiral CT, it has a good guiding significance for defining clinical diagnosis and treatment.

scholarly journals Matching Large Baseline Oblique Stereo Images Using an End-to-End Convolutional Neural Network

2021 ◽  
Vol 13 (2) ◽  
pp. 274
Author(s):  
Guobiao Yao ◽  
Alper Yilmaz ◽  
Li Zhang ◽  
Fei Meng ◽  
Haibin Ai ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Convolutional Neural Network ◽  
Stereo Matching ◽  
Least Square ◽  
Affine Invariant ◽  
Stereo Images ◽  
Distance Ratio ◽  
Matching Algorithm ◽  
End To End

The available stereo matching algorithms produce large number of false positive matches or only produce a few true-positives across oblique stereo images with large baseline. This undesired result happens due to the complex perspective deformation and radiometric distortion across the images. To address this problem, we propose a novel affine invariant feature matching algorithm with subpixel accuracy based on an end-to-end convolutional neural network (CNN). In our method, we adopt and modify a Hessian affine network, which we refer to as IHesAffNet, to obtain affine invariant Hessian regions using deep learning framework. To improve the correlation between corresponding features, we introduce an empirical weighted loss function (EWLF) based on the negative samples using K nearest neighbors, and then generate deep learning-based descriptors with high discrimination that is realized with our multiple hard network structure (MTHardNets). Following this step, the conjugate features are produced by using the Euclidean distance ratio as the matching metric, and the accuracy of matches are optimized through the deep learning transform based least square matching (DLT-LSM). Finally, experiments on Large baseline oblique stereo images acquired by ground close-range and unmanned aerial vehicle (UAV) verify the effectiveness of the proposed approach, and comprehensive comparisons demonstrate that our matching algorithm outperforms the state-of-art methods in terms of accuracy, distribution and correct ratio. The main contributions of this article are: (i) our proposed MTHardNets can generate high quality descriptors; and (ii) the IHesAffNet can produce substantial affine invariant corresponding features with reliable transform parameters.

Convergence and stability analysis of the half thresholding based few-view CT reconstruction

2020 ◽  
Vol 28 (6) ◽  
pp. 829-847
Author(s):  
Hua Huang ◽  
Chengwu Lu ◽  
Lingli Zhang ◽  
Weiwei Wang
Keyword(s):  
Noise Suppression ◽  
Reconstructed Image ◽  
Reference Image ◽  
Projection Data ◽  
Ct Reconstruction ◽  
Reconstruction Algorithms ◽  
Ill Posed ◽  
Thresholding Algorithm ◽  
The Stability ◽  
Well Posed

AbstractThe projection data obtained using the computed tomography (CT) technique are often incomplete and inconsistent owing to the radiation exposure and practical environment of the CT process, which may lead to a few-view reconstruction problem. Reconstructing an object from few projection views is often an ill-posed inverse problem. To solve such problems, regularization is an effective technique, in which the ill-posed problem is approximated considering a family of neighboring well-posed problems. In this study, we considered the {\ell_{1/2}} regularization to solve such ill-posed problems. Subsequently, the half thresholding algorithm was employed to solve the {\ell_{1/2}} regularization-based problem. The convergence analysis of the proposed method was performed, and the error bound between the reference image and reconstructed image was clarified. Finally, the stability of the proposed method was analyzed. The result of numerical experiments demonstrated that the proposed method can outperform the classical reconstruction algorithms in terms of noise suppression and preserving the details of the reconstructed image.

scholarly journals Deep learning for end-to-end kidney cancer diagnosis on multi-phase abdominal computed tomography

2021 ◽  
Vol 5 (1) ◽  
Author(s):  
Kwang-Hyun Uhm ◽  
Seung-Won Jung ◽  
Moon Hyung Choi ◽  
Hong-Kyu Shin ◽  
Jae-Ik Yoo ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Deep Learning ◽  
Kidney Cancer ◽  
Renal Tumor ◽  
Renal Tumors ◽  
Abdominal Computed Tomography ◽  
Wide Range ◽  
End To End ◽  
Multi Phase ◽  
Histologic Subtypes

AbstractIn 2020, it is estimated that 73,750 kidney cancer cases were diagnosed, and 14,830 people died from cancer in the United States. Preoperative multi-phase abdominal computed tomography (CT) is often used for detecting lesions and classifying histologic subtypes of renal tumor to avoid unnecessary biopsy or surgery. However, there exists inter-observer variability due to subtle differences in the imaging features of tumor subtypes, which makes decisions on treatment challenging. While deep learning has been recently applied to the automated diagnosis of renal tumor, classification of a wide range of subtype classes has not been sufficiently studied yet. In this paper, we propose an end-to-end deep learning model for the differential diagnosis of five major histologic subtypes of renal tumors including both benign and malignant tumors on multi-phase CT. Our model is a unified framework to simultaneously identify lesions and classify subtypes for the diagnosis without manual intervention. We trained and tested the model using CT data from 308 patients who underwent nephrectomy for renal tumors. The model achieved an area under the curve (AUC) of 0.889, and outperformed radiologists for most subtypes. We further validated the model on an independent dataset of 184 patients from The Cancer Imaging Archive (TCIA). The AUC for this dataset was 0.855, and the model performed comparably to the radiologists. These results indicate that our model can achieve similar or better diagnostic performance than radiologists in differentiating a wide range of renal tumors on multi-phase CT.

scholarly journals End-to-End, Pixel-Wise Vessel-Specific Coronary and Aortic Calcium Detection and Scoring Using Deep Learning

Diagnostics ◽  
2021 ◽  
Vol 11 (2) ◽  
pp. 215
Author(s):  
Gurpreet Singh ◽  
Subhi Al’Aref ◽  
Benjamin Lee ◽  
Jing Lee ◽  
Swee Tan ◽  
...  
Keyword(s):  
Deep Learning ◽  
Coronary Artery ◽  
Clinical Importance ◽  
Learning Model ◽  
Mean Difference ◽  
Interquartile Range ◽  
History Of ◽  
End To End ◽  
Artery Disease ◽  
Deep Learning Model

Conventional scoring and identification methods for coronary artery calcium (CAC) and aortic calcium (AC) result in information loss from the original image and can be time-consuming. In this study, we sought to demonstrate an end-to-end deep learning model as an alternative to the conventional methods. Scans of 377 patients with no history of coronary artery disease (CAD) were obtained and annotated. A deep learning model was trained, tested and validated in a 60:20:20 split. Within the cohort, mean age was 64.2 ± 9.8 years, and 33% were female. Left anterior descending, right coronary artery, left circumflex, triple vessel, and aortic calcifications were present in 74.87%, 55.82%, 57.41%, 46.03%, and 85.41% of patients respectively. An overall Dice score of 0.952 (interquartile range 0.921, 0.981) was achieved. Stratified by subgroups, there was no difference between male (0.948, interquartile range 0.920, 0.981) and female (0.965, interquartile range 0.933, 0.980) patients (p = 0.350), or, between age <65 (0.950, interquartile range 0.913, 0.981) and age ≥65 (0.957, interquartile range 0.930, 0.9778) (p = 0.742). There was good correlation and agreement for CAC prediction (rho = 0.876, p < 0.001), with a mean difference of 11.2% (p = 0.100). AC correlated well (rho = 0.947, p < 0.001), with a mean difference of 9% (p = 0.070). Automated segmentation took approximately 4 s per patient. Taken together, the deep-end learning model was able to robustly identify vessel-specific CAC and AC with high accuracy, and predict Agatston scores that correlated well with manual annotation, facilitating application into areas of research and clinical importance.

Sign in / Sign up

Export Citation Format

Share Document