• Korean Journal of Radiology
• /
• v.21 no.10
• /
• pp.1165-1177
• /
• 2020

Objective: To assess the feasibility of applying a deep learning-based denoising technique to coronary CT angiography (CCTA) along with iterative reconstruction for additional noise reduction. Materials and Methods: We retrospectively enrolled 82 consecutive patients (male:female = 60:22; mean age, 67.0 ± 10.8 years) who had undergone both CCTA and invasive coronary artery angiography from March 2017 to June 2018. All included patients underwent CCTA with iterative reconstruction (ADMIRE level 3, Siemens Healthineers). We developed a deep learning based denoising technique (ClariCT.AI, ClariPI), which was based on a modified U-net type convolutional neural net model designed to predict the possible occurrence of low-dose noise in the originals. Denoised images were obtained by subtracting the predicted noise from the originals. Image noise, CT attenuation, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were objectively calculated. The edge rise distance (ERD) was measured as an indicator of image sharpness. Two blinded readers subjectively graded the image quality using a 5-point scale. Diagnostic performance of the CCTA was evaluated based on the presence or absence of significant stenosis (≥ 50% lumen reduction). Results: Objective image qualities (original vs. denoised: image noise, 67.22 ± 25.74 vs. 52.64 ± 27.40; SNR [left main], 21.91 ± 6.38 vs. 30.35 ± 10.46; CNR [left main], 23.24 ± 6.52 vs. 31.93 ± 10.72; all p < 0.001) and subjective image quality (2.45 ± 0.62 vs. 3.65 ± 0.60, p < 0.001) improved significantly in the denoised images. The average ERDs of the denoised images were significantly smaller than those of originals (0.98 ± 0.08 vs. 0.09 ± 0.08, p < 0.001). With regard to diagnostic accuracy, no significant differences were observed among paired comparisons. Conclusion: Application of the deep learning technique along with iterative reconstruction can enhance the noise reduction performance with a significant improvement in objective and subjective image qualities of CCTA images.

• Smart Structures and Systems
• /
• v.30 no.6
• /
• pp.687-701
• /
• 2022

Inevitable response loss under complex operational conditions significantly affects the integrity and quality of measured data, leading the structural health monitoring (SHM) ineffective. To remedy the impact of data loss, a common way is to transfer the recorded response of available measure point to where the data loss occurred by establishing the response mapping from measured data. However, the current research has yet addressed the structural condition changes afterward and response mapping learning from a small sample. So, this paper proposes a novel data driven structural response reconstruction method based on a sophisticated designed generating adversarial network (UAGAN). Advanced deep learning techniques including U-shaped dense blocks, self-attention and a customized loss function are specialized and embedded in UAGAN to improve the universal and representative features extraction and generalized responses mapping establishment. In numerical validation, UAGAN efficiently and accurately captures the distinguished features of structural response from only 40 training samples of the intact structure. Besides, the established response mapping is universal, which effectively reconstructs responses of the structure suffered up to 10% random stiffness reduction or structural damage. In the experimental validation, UAGAN is trained with ambient response and applied to reconstruct response measured under earthquake. The reconstruction losses of response in the time and frequency domains reached 16% and 17%, that is better than the previous research, demonstrating the leading performance of the sophisticated designed network. In addition, the identified modal parameters from reconstructed and the corresponding true responses are highly consistent indicates that the proposed UAGAN is very potential to be applied to practical civil engineering.

• Journal of the Korean Society of Radiology
• /
• v.17 no.3
• /
• pp.297-303
• /
• 2023

Pediatric Computed Tomography (CT) examinations can often result in exam failures or the need for frequent retests due to the difficulty of cooperation from young patients. Deep Learning Image Reconstruction (DLIR) methods offer the potential to obtain diagnostically valuable images while reducing the retest rate in CT examinations of pediatric patients with high radiation sensitivity. In this study, we investigated the possibility of applying DLIR to reduce artifacts caused by respiration or motion and obtain clinically useful images in pediatric chest CT examinations. Retrospective analysis was conducted on chest CT examination data of 43 children under the age of 7 from P Hospital in Gyeongsangnam-do. The images reconstructed using Filtered Back Projection (FBP), Adaptive Statistical Iterative Reconstruction (ASIR-50), and the deep learning algorithm TrueFidelity-Middle (TF-M) were compared. Regions of interest (ROI) were drawn on the right ascending aorta (AA) and back muscle (BM) in contrast-enhanced chest images, and noise (standard deviation, SD) was measured using Hounsfield units (HU) in each image. Statistical analysis was performed using SPSS (ver. 22.0), analyzing the mean values of the three measurements with one-way analysis of variance (ANOVA). The results showed that the SD values for AA were FBP=25.65±3.75, ASIR-50=19.08±3.93, and TF-M=17.05±4.45 (F=66.72, p=0.00), while the SD values for BM were FBP=26.64±3.81, ASIR-50=19.19±3.37, and TF-M=19.87±4.25 (F=49.54, p=0.00). Post-hoc tests revealed significant differences among the three groups. DLIR using TF-M demonstrated significantly lower noise values compared to conventional reconstruction methods. Therefore, the application of the deep learning algorithm TrueFidelity-Middle (TF-M) is expected to be clinically valuable in pediatric chest CT examinations by reducing the degradation of image quality caused by respiration or motion.

• Current Optics and Photonics
• /
• v.8 no.3
• /
• pp.246-258
• /
• 2024

In aerospace and energy engineering, the reconstruction of three-dimensional (3D) temperature distributions is crucial. Traditional methods like algebraic iterative reconstruction and filtered back-projection depend on voxel division for resolution. Our algorithm, blending deep learning with computer graphics rendering, converts 2D projections into light rays for uniform sampling, using a fully connected neural network to depict the 3D temperature field. Although effective in capturing internal details, it demands multiple cameras for varied angle projections, increasing cost and computational needs. We assess the impact of camera number on reconstruction accuracy and efficiency, conducting butane-flame simulations with different camera setups (6 to 18 cameras). The results show improved accuracy with more cameras, with 12 cameras achieving optimal computational efficiency (1.263) and low error rates. Verification experiments with 9, 12, and 15 cameras, using thermocouples, confirm that the 12-camera setup as the best, balancing efficiency and accuracy. This offers a feasible, cost-effective solution for real-world applications like engine testing and environmental monitoring, improving accuracy and resource management in temperature measurement.

• Journal of the Korean Society of Radiology
• /
• v.17 no.3
• /
• pp.285-295
• /
• 2023

With the development of technology, efforts to reduce the exposure dose received by patients in CT scans are continuing with the development of new reconstruction techniques. Recently, deep learning reconstruction techniques have been developed to overcome the limitations of repetitive reconstruction techniques. This study aims to evaluate the usefulness of images according to reconstruction techniques in pediatric chest CT images. Patient study conducted a study on 85 pediatric patients who underwent chest CT scan at P-Hospital in Gyeongsangnam-do from January 1, 2021 to December 31, 2022. The phantom used in the Phantom Study is the Pediatrics Whole Body Phantom PBU-70. After the test, the images were reconstructed with FBP, ASIR-V (50%) and DLIR (TF-Medium, High), and the images were evaluated by obtaining SNR and CNR values by setting ROI of the same size. As a result, TF-H of deep learning reconstruction techniques had the lowest noise value compared to ASIR-V (50%) and TF-M in all experiments, and SNR and CNR had the highest values. In pediatric chest CT scans, TF images with deep learning reconstruction techniques were less noisy than ASiR-V images with adaptive statistical iterative reconstruction techniques, CNR and SNR were higher, and the quality of images was improved compared to conventional reconstruction techniques.

• Imaging Science in Dentistry
• /
• v.50 no.4
• /
• pp.331-337
• /
• 2020

Purpose: As cone-beam computed tomography (CBCT) has become the most widely used 3-dimensional (3D) imaging modality in the dental field, storage space and costs for large-capacity data have become an important issue. Therefore, if 3D data can be stored at a clinically acceptable compression rate, the burden in terms of storage space and cost can be reduced and data can be managed more efficiently. In this study, a deep learning network for super-resolution was tested to restore compressed virtual CBCT images. Materials and Methods: Virtual CBCT image data were created with a publicly available online dataset (CQ500) of multidetector computed tomography images using CBCT reconstruction software (TIGRE). A very deep super-resolution (VDSR) network was trained to restore high-resolution virtual CBCT images from the low-resolution virtual CBCT images. Results: The images reconstructed by VDSR showed better image quality than bicubic interpolation in restored images at various scale ratios. The highest scale ratio with clinically acceptable reconstruction accuracy using VDSR was 2.1. Conclusion: VDSR showed promising restoration accuracy in this study. In the future, it will be necessary to experiment with new deep learning algorithms and large-scale data for clinical application of this technology.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.98-101
• /
• 2019

At present, deep convolutional neural networks have made a very important contribution in single-image super-resolution. Through the learning of the neural networks, the features of input images are transformed and combined to establish a nonlinear mapping of low-resolution images to high-resolution images. Some previous methods are difficult to train and take up a lot of memory. In this paper, we proposed a simple and compact deep recursive residual network learning the features for single image super resolution. Global residual learning and local residual learning are used to reduce the problems of training deep neural networks. And the recursive structure controls the number of parameters to save memory. Experimental results show that the proposed method improved image qualities that occur in previous methods.

• Nuclear Engineering and Technology
• /
• v.54 no.5
• /
• pp.1747-1753
• /
• 2022

Compton imaging is the main method for locating radioactive hot spots emitting high-energy gamma-ray photons. In particular, this imaging method is crucial when the photon energy is too high for coded-mask aperture imaging methods to be effective or when a large field of view is required. Reconstruction of the photon source requires advanced Compton event processing algorithms to determine the exact position of the source. In this study, we introduce a novel method based on a Deep Learning algorithm with a Convolutional Neural Network (CNN) to perform Compton imaging. This algorithm is trained on simulated data and tested on real data acquired with Caliste, a single planar CdTe pixelated detector. We show that performance in terms of source location accuracy is equivalent to state-of-the-art algorithms, while computation time is significantly reduced and sensitivity is improved by a factor of ~5 in the Caliste configuration.

• Journal of the Korean Society of Radiology
• /
• v.17 no.1
• /
• pp.37-46
• /
• 2023

This study aims to evaluate the usefulness of Deep Learning Image Reconstruction (TrueFidelity, TF), the image quality of existing Filtered Back Projection (FBP) and Adaptive Statistical Iterative Reconstruction-Veo (ASIR-V) were compared. Noise, CNR, and SSIM were measured by obtaining images with doses fixed at 17.29 mGy and altered to 10.37 mGy, 12.10 mGy, 13.83 mGy, and 15.56 mGy in reconstruction techniques of FBP, ASIR-V 50%, and TF-H. TF-H has superior image quality compared to FBP and ASIR-V when the reconstruction technique change is given at 17.29 mGy. When dose changes were made, Noise, CNR, and SSIM were significantly different when comparing 10.37 mGy TF-H and FBP (p<0.05), and no significant difference when comparing 10.37 mGy TF-H and ASIR-V 50% (p>0.05). TF-H has a dose-reduction effect of 30%, as the highest dose of 15.56 mGy ASIR-V has the same image quality as the lowest dose of 10.37 mGy TF-H. Thus, Deep Learning Reconstruction techniques (TF) were able to reduce dose compared to Iterative Reconstruction techniques (ASIR-V) and Filtered Back Projection (FBP). Therefore, it is considered to reduce the exposure dose of patients.

• Korean Journal of Remote Sensing
• /
• v.39 no.1
• /
• pp.1-21
• /
• 2023

In this technically advanced era, we are surrounded by smartphones, computers, and cameras, which help us to store visual information in 2D image planes. However, such images lack 3D spatial information about the scene, which is very useful for scientists, surveyors, engineers, and even robots. To tackle such problems, depth maps are generated for respective image planes. Depth maps or depth images are single image metric which carries the information in three-dimensional axes, i.e., xyz coordinates, where z is the object's distance from camera axes. For many applications, including augmented reality, object tracking, segmentation, scene reconstruction, distance measurement, autonomous navigation, and autonomous driving, depth estimation is a fundamental task. Much of the work has been done to calculate depth maps. We reviewed the status of depth map estimation using different techniques from several papers, study areas, and models applied over the last 20 years. We surveyed different depth-mapping techniques based on traditional ways and newly developed deep-learning methods. The primary purpose of this study is to present a detailed review of the state-of-the-art traditional depth mapping techniques and recent deep learning methodologies. This study encompasses the critical points of each method from different perspectives, like datasets, procedures performed, types of algorithms, loss functions, and well-known evaluation metrics. Similarly, this paper also discusses the subdomains in each method, like supervised, unsupervised, and semi-supervised methods. We also elaborate on the challenges of different methods. At the conclusion of this study, we discussed new ideas for future research and studies in depth map research.