• Journal of IKEEE
• /
• v.28 no.2
• /
• pp.234-238
• /
• 2024

This paper presents a study on the integration of 3D-stacked dual-tip RRAM with a CNN accelerator architecture, leveraging its low drive current characteristics and scalability in a 3D stacked configuration. The dual-tip structure is utilized in a parallel connection format in a synaptic array to implement multi-level capabilities. It is configured within a Network-on-chip style accelerator along with various hardware blocks such as DAC, ADC, buffers, registers, and shift & add circuits, and simulations were performed for the CNN accelerator. The quantization of synaptic weights and activation functions was assumed to be 16-bit. Simulation results of CNN operations through a parallel pipeline for this accelerator architecture achieved an operational efficiency of approximately 370 GOPs/W, with accuracy degradation due to quantization kept within 3%.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.113-118
• /
• 2024

In this paper, we propose an optimal wavelet in a system for canceling background noise of acoustic signals. This system performed Discrete Wavelet Transform(DWT) instead of the existing Short Time Fourier Transform(STFT) and then improved noise cancellation performance through a deep learning process. DWT functions as a multi-resolution band-pass filter and obtains transformation parameters by time-shifting the parent wavelet at each level and using several wavelets whose sizes are scaled. Here, the noise cancellation performance of several wavelets was tested to select the most suitable mother wavelet for analyzing the speech. In this study, to verify the performance of the noise cancellation system for various wavelets, a simulation program using Tensorflow and Keras libraries was created and simulation experiments were performed for the four most commonly used wavelets. As a result of the experiment, the case of using Haar or Daubechies wavelets showed the best noise cancellation performance, and the mean square error(MSE) was significantly improved compared to the case of using other wavelets.

• Imaging Science in Dentistry
• /
• v.54 no.2
• /
• pp.191-199
• /
• 2024

Purpose: The aim of this study was to evaluate image artifacts in the vicinity of dental implants in cone-beam computed tomography (CBCT) scans obtained with different spatial orientations, tube current levels, and metal artifact reduction algorithm (MAR) conditions. Materials and Methods: One dental implant and 2 tubes filled with a radiopaque solution were placed in the posterior region of a mandible using a surgical guide to ensure parallel alignment. CBCT scans were acquired with the mandible in 2 spatial orientations in relation to the X-ray projection plane (standard and modified) at 3 tube current levels: 5, 8, and 11 mA. CBCT scans were repeated without the implant and were reconstructed with and without MAR. The mean voxel and noise values of each tube were obtained and compared using multi-way analysis of variance and the Tukey test(α=0.05). Results: Mean voxel values were significantly higher and noise values were significantly lower in the modified orientation than in the standard orientation (P<0.05). MAR activation and tube current levels did not show significant differences in most cases of the modified spatial orientation and in the absence of the dental implant (P>0.05). Conclusion: Modifying the spatial orientation of the head increased brightness and reduced spatial orientation noise in adjacent regions of a dental implant, with no influence from the tube current level and MAR.

• Clinical Nutrition Research
• /
• v.12 no.4
• /
• pp.269-282
• /
• 2023

Vitamin D participates in the biological function of the innate and adaptive immune system and inflammation. We aim to specify the effectiveness of the vitamin D supplementation on the side effects BioNTech, Pfizer vaccination, and immunoglobulin G response against severe acute respiratory syndrome coronavirus 2 in subjects tested positive for coronavirus disease 2019 (COVID-19). In this multi-center randomized clinical trial, 498 people tested positive for COVID-19 were divided into 2 groups, receiving vitamin D capsules or a placebo (1 capsule daily, each containing 600 IU of vitamin D) over 14-16 weeks. Anthropometric indices and biochemical parameters were measured before and after the second dose of vaccination. Fourteen to 16 weeks after supplementation, the intervention group had an immunoglobulin G (IgG) increase of 10.89 ± 1.2 g/L, while the control group had 8.89 ± 1.3 g/L, and the difference was significant between both groups (p = 0.001). After the second dose of vaccination, the supplement group significantly increased their 25-hydroxy vitamin D from initially 28.73 ± 15.6 ng/mL and increased to 46.48 ± 27.2 ng/mL, and the difference between them was significant. Those with a higher body mass index (BMI) had the most of symptoms, and the difference of side effects according to BMI level was significantly different. In 8 weeks after supplementation obese participants had the lowest IgG levels than overweight or normal subjects. The proportion of all types of side effects on the second dose was significantly diminished compared with the first dose in the intervention group. Supplementation of 600 IU of vitamin D3 can reduce post-vaccination side effects and increase IgG levels in participants who received BioNTech, Pfizer vaccine.

• Journal of Radiation Protection and Research
• /
• v.49 no.1
• /
• pp.29-39
• /
• 2024

Background: The gamma emission tomography (GET) device has been reported a reliable technique to inspect partial defects within spent nuclear fuel (SNF) of pin-by-pin level. However, the existing GET devices have low accuracy owing to the high attenuation and scatter probability for SNF inspection condition. The purpose of this study is to design and optimize a Yonsei single-photon emission computed tomography version 2 (YSECT.v.2) for fast inspection of SNF in water storage by acquisition of high-quality tomographic images. Materials and Methods: Using Geant4 (Geant4 Collaboration) and DETECT-2000 (Glenn F. Knoll et al.) Monte Carlo simulation, the geometrical structure of the proposed device was determined and its performance was evaluated for the ¹³⁷Cs source in water. In a Geant4-based assessment, proposed device was compared with the International Atomic Energy Agency (IAEA)-authenticated device for the quality of tomographic images obtained for 12 fuel sources in a 14 × 14 Westinghouse-type fuel assembly. Results and Discussion: According to the results, the length, slit width, and septal width of the collimator were determined to be 65, 2.1, and 1.5 mm, respectively, and the material and length of the trapezoidal-shaped scintillator were determined to be gadolinium aluminum gallium garnet and 45 mm, respectively. Based on the results of performance comparison between the YSECT.v.2 and IAEA's device, the proposed device showed 200 times higher performance in gamma-detection sensitivity and similar source discrimination probability. Conclusion: In this study, we optimally designed the GET device for improving the SNF inspection accuracy and evaluated its performance. Our results show that the YSECT.v.2 device could be employed for SNF inspection.

• International Journal of Computer Science & Network Security
• /
• v.23 no.12
• /
• pp.151-160
• /
• 2023

Breast cancer is the most dangerous and deadly form of cancer. Initial detection of breast cancer can significantly improve treatment effectiveness. The second most common cancer among Indian women in rural areas. Early detection of symptoms and signs is the most important technique to effectively treat breast cancer, as it enhances the odds of receiving an earlier, more specialist care. As a result, it has the possible to significantly improve survival odds by delaying or entirely eliminating cancer. Mammography is a high-resolution radiography technique that is an important factor in avoiding and diagnosing cancer at an early stage. Automatic segmentation of the breast part using Mammography pictures can help reduce the area available for cancer search while also saving time and effort compared to manual segmentation. Autoencoder-like convolutional and deconvolutional neural networks (CN-DCNN) were utilised in previous studies to automatically segment the breast area in Mammography pictures. We present Automatic SegmenAN, a unique end-to-end adversarial neural network for the job of medical image segmentation, in this paper. Because image segmentation necessitates extensive, pixel-level labelling, a standard GAN's discriminator's single scalar real/fake output may be inefficient in providing steady and appropriate gradient feedback to the networks. Instead of utilising a fully convolutional neural network as the segmentor, we suggested a new adversarial critic network with a multi-scale L1 loss function to force the critic and segmentor to learn both global and local attributes that collect long- and short-range spatial relations among pixels. We demonstrate that an Automatic SegmenAN perspective is more up to date and reliable for segmentation tasks than the state-of-the-art U-net segmentation technique.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.3
• /
• pp.87-94
• /
• 2024

We investigated the feasibility of utilizing Large Language Models (LLMs) to perform text-based games without training on game data in advance. We adopted ChatGPT-3.5 and its state-of-the-art, ChatGPT-4, as the systems that implemented LLM. In addition, we added the persistent memory feature proposed in this paper to ChatGPT-4 to create three game player agents. We used Zork, one of the most famous text-based games, to see if the agents could navigate through complex locations, gather information, and solve puzzles. The results showed that the agent with persistent memory had the widest range of exploration and the best score among the three agents. However, all three agents were limited in solving puzzles, indicating that LLM is vulnerable to problems that require multi-level reasoning. Nevertheless, the proposed agent was still able to visit 37.3% of the total locations and collect all the items in the locations it visited, demonstrating the potential of LLM.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.3
• /
• pp.659-666
• /
• 2024

Germaine Tailleferre is the only woman composer among the French group of six composers known as "Les Six." In her 70-year career, she has left behind numerous chamber music pieces for the piano. Although her chamber music works constitute a significant portion of her overall compositions, research focusing on her piano chamber music pieces is lacking. Therefore, this study introduces a comprehensive list of Tailleferre's chamber music pieces and categorizes each piece according to its performing level of difficulty. Additionally, through a detailed analysis of her <Sonata for Two Pianos>, composed in 1974, this study aims to understand her musical style and artistic world, particularly regarding form, harmony, and melody. <Sonata for Two Pianos>, rooted in the unpretentious and light musical language characteristic of the salon style popular in Parisian cafes and music halls at the time, can be seen as a multi-layered work reflecting various musical languages such as Impressionism, and Neo-classicism. This study may contribute to a better understanding of Tailleferre's musical world and aid in discovering and expanding new literature on 20th-century piano chamber music.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.44 no.3
• /
• pp.329-336
• /
• 2024

In the case of multi-function weirs installed in Korea, the free hydraulic jump or the submerged hydraulic jump is occurred depending on the height of the gate opening and the tailwater level when the sluice gate of the movable weir is partially opened. In this study, the submerged hydraulic jump for the flows under the sluice gate were simulated and the mean flow, turbulence statistics, and relative water depth are investigated using numerical simulation. For numerical simulation, the unsteady Reynolds-averaged Navier-Stokes equation, volume of fluid method, and k-ωSST turbulence model were used. The numerical model was validated using the results of other researchers' previously performed experiments, and it was investigated that the numerical model appropriately simulates the two-phase flow in the hydraulic jump. In addition, the distribution of mean flow, turbulence statistics, and the length of recirculation region was investigated.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• v.50 no.4
• /
• pp.216-221
• /
• 2024

Objectives: This study aimed to develop and validate a model to predict the need for intensive care unit (ICU) admission in patients with dental infections using an automated machine learning (ML) program called H2O-AutoML. Materials and Methods: Two models were created using only the information available at the initial examination. Model 1 was parameterized with only clinical symptoms and blood tests, excluding contrast-enhanced multi-detector computed tomography (MDCT) images available at the initial visit, whereas model 2 was created with the addition of the MDCT information to the model 1 parameters. Although model 2 was expected to be superior to model 1, we wanted to independently determine this conclusion. A total of 210 patients who visited the Department of Oral and Maxillofacial Surgery at the Dankook University Dental Hospital from March 2013 to August 2023 was included in this study. The patients' demographic characteristics (sex, age, and place of residence), systemic factors (hypertension, diabetes mellitus [DM], kidney disease, liver disease, heart disease, anticoagulation therapy, and osteoporosis), local factors (smoking status, site of infection, postoperative wound infection, dysphagia, odynophagia, and trismus), and factors known from initial blood tests were obtained from their medical charts and retrospectively reviewed. Results: The generalized linear model algorithm provided the best diagnostic accuracy, with an area under the receiver operating characteristic values of 0.8289 in model 1 and 0.8415 in model 2. In both models, the C-reactive protein level was the most important variable, followed by DM. Conclusion: This study provides unprecedented data on the use of ML for successful prediction of ICU admission based on initial examination results. These findings will considerably contribute to the development of the field of dentistry, especially oral and maxillofacial surgery.