• The korean journal of orthodontics
• /
• v.54 no.1
• /
• pp.48-58
• /
• 2024

Objective: To quantify the effects of midline-related landmark identification on midline deviation measurements in posteroanterior (PA) cephalograms using a cascaded convolutional neural network (CNN). Methods: A total of 2,903 PA cephalogram images obtained from 9 university hospitals were divided into training, internal validation, and test sets (n = 2,150, 376, and 377). As the gold standard, 2 orthodontic professors marked the bilateral landmarks, including the frontozygomatic suture point and latero-orbitale (LO), and the midline landmarks, including the crista galli, anterior nasal spine (ANS), upper dental midpoint (UDM), lower dental midpoint (LDM), and menton (Me). For the test, Examiner-1 and Examiner-2 (3-year and 1-year orthodontic residents) and the Cascaded-CNN models marked the landmarks. After point-to-point errors of landmark identification, the successful detection rate (SDR) and distance and direction of the midline landmark deviation from the midsagittal line (ANS-mid, UDM-mid, LDM-mid, and Me-mid) were measured, and statistical analysis was performed. Results: The cascaded-CNN algorithm showed a clinically acceptable level of point-to-point error (1.26 mm vs. 1.57 mm in Examiner-1 and 1.75 mm in Examiner-2). The average SDR within the 2 mm range was 83.2%, with high accuracy at the LO (right, 96.9%; left, 97.1%), and UDM (96.9%). The absolute measurement errors were less than 1 mm for ANS-mid, UDM-mid, and LDM-mid compared with the gold standard. Conclusions: The cascaded-CNN model may be considered an effective tool for the auto-identification of midline landmarks and quantification of midline deviation in PA cephalograms of adult patients, regardless of variations in the image acquisition method.

• Journal of the Korean Society of Radiology
• /
• v.18 no.2
• /
• pp.135-144
• /
• 2024

Ultrasonography examination has limitations in quantifying hepatic fat quantification. Therefore, this study aimed to experimentally demonstrate whether changes in signal attenuation during ultrasound imaging can be quantified using simulated hepatic phantoms to assess hepatic fat content. Additionally, we aimed to evaluate the potential of ultrasound imaging for diagnosing hepatic fatty liver by analyzing the relationship between hepatic fat content and signal intensity in ultrasound images. In this study, we developed a total of five stimulated hepatic phantoms by homogeneously mixing water and oil. We confirmed the fat content of the phantoms using magnetic resonance imaging (MRI) and ultrasound imaging, and measured signal intensity according to distance in ultrasound images to analyze the correlation and mean comparison between fat content and signal intensity. We observed that as the fat content increased, the ultrasound penetration intensity decreased, confirming the potential for quantifying hepatic fat content using ultrasound. Additionally, the analysis of the correlation between the measured fat content using MRI and the signal intensity measured in ultrasound images showed a high correlation. Statistical analysis in our study confirmed that as the fat content increased, the slope representing signal during ultrasound imaging (US-GRE) decreased. In this study, it was statistically confirmed that the US-GRE value of ultrasound images gradually decreases as the fat content increases, and it is believed that US-GRE can serve as a biomarker expressing fatty liver content.

• Journal of radiological science and technology
• /
• v.47 no.3
• /
• pp.205-211
• /
• 2024

Currently, with the development of technologies, X-ray examinations for medical examinations at hospital is increasing. This study was conducted to help reduce radiation exposure by measuring the exposure dose received by pediatric patients and the spatial dose of the X-ray room. Dosimeters were installed in the eyeball, thyroid gland, breast, gonads and 4 directions at a distance of 30 cm, 40 cm, 50 cm from the phantom. The dose was measured ten times each, before, and after the application of the bismuth shield under the examination conditions of the head, chest, and abdomen of pediatric patients. Under the condition of head examination, when a shielding was applied, the dose reduction rate was 68.58% for the eyeball, 72.88% for the thyroid, 84.2% for the breast, and 72.36% for the gonad. The chest examination showed reductions of 19.56% eyeball, 56.98% thyroid, 1.21% breast, and 0.68% gonad. The abdominal examination showed reduction rates of 2.6% eyeball, 10.67% thyroid, 19.85% breast, and 82.02% gonad. Spatial dose decreased by 62.25% at 30 cm, 61.16% at 40 cm, and 68.68% at 50 cm. When the bismuth shield was applied, there was a decrease in dose across all examinations, as well as a reduction in spatial dose. Continued research on the use of bismuth shields will help radiological technologists achieve their goal of dose reduction.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.7
• /
• pp.11-20
• /
• 2024

In this paper, we propose the BAG (Bit-width Aware Generator) and the Intermediate Layer Knowledge Distillation using Channel-wise Attention to reduce the knowledge gap between a quantized network, a full-precision network, and a generator in GDFQ (Generative Data-Free Quantization). Since the generator in GDFQ is only trained by the feedback from the full-precision network, the gap resulting in decreased capability due to low bit-width of the quantized network has no effect on training the generator. To alleviate this problem, BAG is quantized with same bit-width of the quantized network, and it can generate synthetic images, which are effectively used for training the quantized network. Typically, the knowledge gap between the quantized network and the full-precision network is also important. To resolve this, we compute channel-wise attention of outputs of convolutional layers, and minimize the loss function as the distance of them. As the result, the quantized network can learn which channels to focus on more from mimicking the full-precision network. To prove the efficiency of proposed methods, we quantize the network trained on CIFAR-100 with 3 bit-width weights and activations, and train it and the generator with our method. As the result, we achieve 56.14% Top-1 Accuracy and increase 3.4% higher accuracy compared to our baseline AdaDFQ.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.28 no.1
• /
• pp.98-106
• /
• 2024

Due to advancements in construction technology and analytical tools, an increasing number of cable-stayed bridges have been designed and constructed in recent years. A cable is a structural element that primarily transmits the main load of a cable-stayed bridge and plays the most crucial role in reflecting the overall condition of the entire bridge system. In this study, a vision-based method was applied to estimate the tension of the stay cables located at a long distance. To measure the response of a cable using a vision-based method, it is necessary to install feature points or targets on the cable. However, depending on the location of the point to be measured, there may be no feature points in the cable, and there may also be limitations in installing the target on the cable. Hence, it is necessary to find a way to measure cable response that overcomes the limitations of existing vision-based methods. This study proposes a method for measuring cable responses by utilizing the characteristics of cable shape. The proposed method involved extracting the cable shape from the acquired image and determining the center of the extracted cable shape to measure the cable response. The extracted natural frequencies of the vibration mode were obtained using the measured responses, and the tension was estimated by applying them to the vibration method. To verify the reliability of the vision-based method, cable images were obtained from the Hwatae Bridge in service under ambient vibration conditions. The reliability of the method proposed in this study was confirmed by applying it to the vibration method using a vision-based approach, resulting in estimated tensions with an error of less than 1% compared to tensions estimated using an accelerometer.

• Journal of Korean Society of Disaster and Security
• /
• v.17 no.1
• /
• pp.1-8
• /
• 2024

In Korea, mountainous areas cover 60% of the land, leading to increased factors such as concentrated heavy rainfall and typhoons, which can result in debris flow and landslide. Despite the high risk of disasters like landslides and debris flow, there has been a tendency in most regions to focus more on post-damage recovery rather than preventing damage. Therefore, in this study, precise topographic data was constructed by conducting on-site surveys and drone measurements in areas where debris flow actually occurred, to analyze the risk zones for such events. The numerical analysis program RAMMS model was utilized to perform debris flow analysis on the areas prone to debris flow, and the actual distribution of debris flow was compared and analyzed to evaluate the applicability of the model. As a result, the debris flow generation area calculated by the RAMMS model was found to be 18% larger than the actual area, and the travel distance was estimated to be 10% smaller. However, the simulated shape of debris flow generation and the path of movement calculated by the model closely resembled the actual data. In the future, we aim to conduct additional research, including model verification suitable for domestic conditions and the selection of areas for damage prediction through debris flow analysis in unmeasured watersheds.

• Korean Journal of Radiology
• /
• v.24 no.4
• /
• pp.349-361
• /
• 2023

Objective: To quantitatively assess the pulmonary vasculature using non-contrast computed tomography (CT) in patients with chronic thromboembolic pulmonary hypertension (CTEPH) pre- and post-treatment and correlate CT-based parameters with right heart catheterization (RHC) hemodynamic and clinical parameters. Materials and Methods: A total of 30 patients with CTEPH (mean age, 57.9 years; 53% female) who received multimodal treatment, including riociguat for ≥ 16 weeks with or without balloon pulmonary angioplasty and underwent both non-contrast CT for pulmonary vasculature analysis and RHC pre- and post-treatment were included. The radiographic analysis included subpleural perfusion parameters, including blood volume in small vessels with a cross-sectional area ≤ 5 mm² (BV5) and total blood vessel volume (TBV) in the lungs. The RHC parameters included mean pulmonary artery pressure (mPAP), pulmonary vascular resistance (PVR), and cardiac index (CI). Clinical parameters included the World Health Organization (WHO) functional class and 6-minute walking distance (6MWD). Results: The number, area, and density of the subpleural small vessels increased after treatment by 35.7% (P < 0.001), 13.3% (P = 0.028), and 39.3% (P < 0.001), respectively. The blood volume shifted from larger to smaller vessels, as indicated by an 11.3% increase in the BV5/TBV ratio (P = 0.042). The BV5/TBV ratio was negatively correlated with PVR (r = -0.26; P = 0.035) and positively correlated with CI (r = 0.33; P = 0.009). The percent change across treatment in the BV5/TBV ratio correlated with the percent change in mPAP (r = -0.56; P = 0.001), PVR (r = -0.64; P < 0.001), and CI (r = 0.28; P = 0.049). Furthermore, the BV5/TBV ratio was inversely associated with the WHO functional classes I-IV (P = 0.004) and positively associated with 6MWD (P = 0.013). Conclusion: Non-contrast CT measures could quantitatively assess changes in the pulmonary vasculature in response to treatment and were correlated with hemodynamic and clinical parameters.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.34 no.3
• /
• pp.79-85
• /
• 2024

We investigated the crystal structural property and chemical bonding nature of cellulose nanocrystal extracted directly from cotton cellulose using high-pressure homogenizer. The nanowire-like cellulose nanocrystals were randomly distributed in the form of a dense mesh. Based on calculating the interplanar distance of the Bragg-diffracted crystal plane observed through X-ray diffraction (XRD) analysis, it was found that the cellulose nanocrystals formed by high-pressure homogenizer had a monoclinc crystal structure, corresponding to the cellulose Iβ sub-polymorph. Solid-state nuclear magnetic resonance (NMR) analysis for the quantitatively evaluation of the amorphous region in cellulose nanocrystals revealed that the crystallinity index of cellulose nanocrystals was calculated to be 53.06 %. The O/C ratio of the surface of cellulose nanocrystal was estimated to be 0.82. Further analysis showed that chemical bonds of C-C bond or C-H bond, C-O bond, O-C-O bond or C=O bond, and O-C=O bond were the main chemical bonding states of the cellulose nanocrystal surface.

• Korean Journal of Radiology
• /
• v.23 no.11
• /
• pp.1044-1054
• /
• 2022

Objective: This study aimed to investigate whether a deep learning reconstruction (DLR) method improves the image quality, stent evaluation, and visibility of the valve apparatus in coronary computed tomography angiography (CCTA) when compared with filtered back projection (FBP) and hybrid iterative reconstruction (IR) methods. Materials and Methods: CCTA images of 51 patients (mean age ± standard deviation [SD], 63.9 ± 9.8 years, 36 male) who underwent examination at a single institution were reconstructed using DLR, FBP, and hybrid IR methods and reviewed. CT attenuation, image noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and stent evaluation, including 10%-90% edge rise slope (ERS) and 10%-90% edge rise distance (ERD), were measured. Quantitative data are summarized as the mean ± SD. The subjective visual scores (1 for worst -5 for best) of the images were obtained for the following: overall image quality, image noise, and appearance of stent, vessel, and aortic and tricuspid valve apparatus (annulus, leaflets, papillary muscles, and chordae tendineae). These parameters were compared between the DLR, FBP, and hybrid IR methods. Results: DLR provided higher Hounsfield unit (HU) values in the aorta and similar attenuation in the fat and muscle compared with FBP and hybrid IR. The image noise in HU was significantly lower in DLR (12.6 ± 2.2) than in hybrid IR (24.2 ± 3.0) and FBP (54.2 ± 9.5) (p < 0.001). The SNR and CNR were significantly higher in the DLR group than in the FBP and hybrid IR groups (p < 0.001). In the coronary stent, the mean value of ERS was significantly higher in DLR (1260.4 ± 242.5 HU/mm) than that of FBP (801.9 ± 170.7 HU/mm) and hybrid IR (641.9 ± 112.0 HU/mm). The mean value of ERD was measured as 0.8 ± 0.1 mm for DLR while it was 1.1 ± 0.2 mm for FBP and 1.1 ± 0.2 mm for hybrid IR. The subjective visual scores were higher in the DLR than in the images reconstructed with FBP and hybrid IR. Conclusion: DLR reconstruction provided better images than FBP and hybrid IR reconstruction.

• Information Systems Review
• /
• v.21 no.3
• /
• pp.111-129
• /
• 2019

Recently, artificial intelligence (AI)-enabled products and services such as smartphones, smart speakers, chatbots are being released due to advances in AI technology. Thus researchers making effort to reveal that consumers' intention to adopt AI-enabled products. Yet, little is known about the intended adoption of AI-enabled products. Because most of studies has been not consideredthe perceived utility value of consumers for each attribute by classified based on the characteristics of AI-enabled products. Therefore, the purpose of this study is to investigate the difference in importance between attributes that affect the intention to adopt of AI-enabled products. For this, first, identified and classified the attributes of AI-enabled products based on IS Success Model of DeLone and McLean. Second, measured the utility value of each attribute on the adoption of AI-enabled products through conjoint analysis. And we employed construal level theory to see whether there are differences in the relative importance of AI-enabled products attributes depending on the temporal distance. Third, we segmented the market based on the utility value of each respondent through cluster analysis and tried to understand the characteristics and needs of consumers in each segment market. We expect to provide theoretical implications for conceptually structured attributes and factors of AI-enabled products and practical implications for how development efforts of AI-enabled products are needed to reach consumers need for each segment.