• Journal of the Korea Society of Computer and Information
• /
• v.27 no.5
• /
• pp.77-84
• /
• 2022

In this paper, we propose a novel unsupervised feature selection method. Conventional unsupervised feature selection method defines virtual label and uses a regression analysis that projects the given data to this label. However, since virtual labels are generated from data, they can be formed similarly in the space. Thus, in the conventional method, the features can be selected in only restricted space. To solve this problem, in this paper, features are selected using orthogonal projections and low-rank approximations. To solve this problem, in this paper, a virtual label is projected to orthogonal space and the given data set is also projected to this space. Through this process, effective features can be selected. In addition, projection matrix is restricted low-rank to allow more effective features to be selected in low-dimensional space. To achieve these objectives, a cost function is designed and an efficient optimization method is proposed. Experimental results for six data sets demonstrate that the proposed method outperforms existing conventional unsupervised feature selection methods in most cases.

• Current Optics and Photonics
• /
• v.6 no.2
• /
• pp.202-211
• /
• 2022

In the pasteurization of escalator handrails using ultraviolet (UV) sterilizers, a combination of light distribution and escalator speed has priority over other important factors. Furthermore, since part of the escalator handrail has a curved structure, proper design is needed to improve the sterilization rate on the surfaces touched by users. In this paper, two types of sterilizers satisfying these conditions are manufactured with 275-nm UV-C LEDs, after modeling the three-dimensional (3D) structure of an escalator handrail and simulating optical distributions of UV-C irradiation on the handrail's surface according to light-emitting diode (LED) positions and reflector variations in the sterilizers. Pasteurization experiments with the UV-C LED sterilizers are conducted on six types of gram-positive and gram-negative bacteria, with exposure times of 0.2, 5, and 15 s at an actual installation distance of 20 mm. The sterilization rates for the gram-positive bacteria are 10.63% to 27.94% at 0.2 s, 89.44% to 96.30% at 5 s, and 99.64% to 99.88% at 15 s. Those for the gram-negative bacteria are 57.70% to 77.63% at 0.2 s, 98.90% to 99.49% at 5 s, and 99.88% to 99.99% at 15 s. The power consumption of the UV-C LED sterilizer is about 8 W, which can be supplied by a self-generation module instead of an external power supply.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.4
• /
• pp.1-8
• /
• 2009

In this paper, we deal an implementation of the software that produces sub-optimal solution of cutting-ironbar planning problem using dynamic programming. Generally, it is required to design an optimization algorithm to accept the practical requirements of cutting ironbar manufacturing. But, this problem is a multiple-sized 1-dimensional cutting stock problem and Linear Programming approaches to get the optimal solution is difficult to be applied due to the problem of explosive computation and memory limitation. In order to overcome this problem, we reform the problem for applying Dynamic Programming and propose a cutting-ironbar planning algorithm searching the sub-optimal solution in the space of fixed amount of combinated columns by using heuristics. Then, we design a graphic user interfaces and screen displays to be operated conveniently in the industry workplace and implement the software using open-source GUI library toolkit, GTK+.

• Nuclear Engineering and Technology
• /
• v.55 no.5
• /
• pp.1587-1592
• /
• 2023

Time-encoded imagers (TEI) are important class of instruments to search for potential radioactive sources to prevent illicit transportation and trafficking of nuclear materials and other radioactive sources. The energy of the radiation cannot be known in advance due to the type and shielding of source is unknown in practice. However, the response function of the time-encoded imagers is related to the energy of neutrons or gamma-rays. An improved image reconstruction method based on MLEM was proposed to correct for the energy induced response difference. In this method, the count vector versus time was first smoothed. Then, the preset response function was adaptively corrected according to the measured counts. Finally, the smoothed count vector and corrected response were used in MLEM to reconstruct the source distribution. A one-dimensional dual-particle time-encode imager was developed and used to verify the improved method through imaging an Am-Be neutron source. The improvement of this method was demonstrated by the image reconstruction results. For gamma-ray and neutron images, the angular resolution improved by 17.2% and 7.0%; the contrast-to-noise ratio improved by 58.7% and 14.9%; the signal-to-noise ratio improved by 36.3% and 11.7%, respectively.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.4
• /
• pp.357-363
• /
• 2023

In underwater signal processing, separating individual signals from mixed signals has long been a challenge due to low signal quality. The common method using Short-time Fourier transform for spectrogram analysis has faced criticism for its complex parameter optimization and loss of phase data. We propose a Triple-path Recurrent Neural Network, based on the Dual-path Recurrent Neural Network's success in long time series signal processing, to handle three-dimensional tensors from multi-channel sensor input signals. By dividing input signals into short chunks and creating a 3D tensor, the method accounts for relationships within and between chunks and channels, enabling local and global feature learning. The proposed technique demonstrates improved Root Mean Square Error and Scale Invariant Signal to Noise Ratio compared to the existing method.

• Imaging Science in Dentistry
• /
• v.53 no.4
• /
• pp.345-353
• /
• 2023

Purpose: The objective of this study was to propose a method for developing a clinical phantom to reproduce various diseases that are clinically prevalent in the field of dentistry. This could facilitate diverse clinical research without unnecessarily exposing patients to radiation. Materials and Methods: This study utilized a single dry skull, which was visually and radiographically examined to evaluate its condition. Existing lesions on the dry skull were preserved, and other relevant lesions were artificially created as necessary. These lesions were then documented using intraoral radiography and cone-beam computed tomography. Once all pre-existing and reproduced lesions were confirmed by the consensus of 2 oral and maxillofacial radiologists, the skull was embedded in a soft tissue substitute. To validate the process, cone-beam computed tomography scans and panoramic radiographs were obtained of the fabricated phantom. All acquired images were subsequently evaluated. Results: Most lesions could be identified on panoramic radiographs, although some sialoliths and cracked teeth were confirmed only through cone-beam computed tomographic images. A small gap was observed between the epoxy resin and the bone structures. However, 2 oral and maxillofacial radiologists agreed that this space did not meaningfully impact the interpretation process. Conclusion: The newly developed phantom has potential for use as a standardized phantom within the dental field. It may be utilized for a variety of imaging studies, not only for optimization purposes, but also for addressing other experimental issues related to both 2- and 3-dimensional diagnostic radiography.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.2
• /
• pp.311-326
• /
• 2024

The rapid development of neural network technology promotes the neural network model driven by big data to overcome the texture effect of complex objects. Due to the limitations in complex scenes, it is necessary to establish custom template matching and apply it to the research of many fields of computational vision technology. The dependence on high-quality small label sample database data is not very strong, and the machine learning system of deep feature connection to complete the task of texture effect inference and speculation is relatively poor. The style transfer algorithm based on neural network collects and preserves the data of patterns, extracts and modernizes their features. Through the algorithm model, it is easier to present the texture color of patterns and display them digitally. In this paper, according to the texture effect reasoning of custom template matching, the 3D visualization of the target is transformed into a 3D model. The high similarity between the scene to be inferred and the user-defined template is calculated by the user-defined template of the multi-dimensional external feature label. The convolutional neural network is adopted to optimize the external area of the object to improve the sampling quality and computational performance of the sample pyramid structure. The results indicate that the proposed algorithm can accurately capture the significant target, achieve more ablation noise, and improve the visualization results. The proposed deep convolutional neural network optimization algorithm has good rapidity, data accuracy and robustness. The proposed algorithm can adapt to the calculation of more task scenes, display the redundant vision-related information of image conversion, enhance the powerful computing power, and further improve the computational efficiency and accuracy of convolutional networks, which has a high research significance for the study of image information conversion.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.45 no.2
• /
• pp.261-268
• /
• 2016

This study was conducted to establish roasting conditions for optimization of Citri Unshii Pericarpium antioxidant activity using response surface methodology (RSM). A central composite design was applied to investigate the effects of two independent variables, namely roasting temperature ($40{\sim}100^{\circ}C$; $X_1$) and roasting time ($5{\sim}15min$; $X_2$), on responses such as electron donating ability ($Y_1$), total phenolic content ($Y_2$), total flavonoid content ($Y_3$), and hydroxyl radical scavenging activity ($Y_4$). The maximum electron donating ability was 72.38% at a roasting temperature of $71.12^{\circ}C$ and roasting time of 9.39 min. The maximum total phenolic content was 10.76 mg tannic acid equivalents/g at a roasting temperature of $69.71^{\circ}C$ and roasting time of 8.39 min. The maximum total flavonoid content was 105.99 mg quercetin equivalents/100 g at $72.54^{\circ}C$ and 8.64 min. The maximum hydroxyl radical scavenging activity was 60.33% at $68.97^{\circ}C$ and 9.84 min. Based on the superimposition of three dimensional RSM with respect to electron donating ability, total phenolic content, total flavonoid content, and hydroxyl radical scavenging activity under various conditions, optimum conditions were established as follows: roasting temperature of $70.90^{\circ}C$ and roasting time of 9.03 min.

• The Journal of Korean Society for Radiation Therapy
• /
• v.20 no.1
• /
• pp.11-15
• /
• 2008

Purpose: The goal of radiation treatment is to deliver a prescribed radiation dose to the target volume accurately while minimizing dose to normal tissues. In this paper, we comparing the dose distribution between three dimensional conformal radiation radiotherapy (3D-CRT) and helical tomotherapy (TOMO) plan for partial breast cancer. Materials and Methods: Twenty patients were included in the study, and plans for two techniques were developed for each patient (left breast:10 patients, right breast:10 patients). For each patient 3D-CRT planning was using pinnacle planning system, inverse plan was made using Tomotherapy Hi-Art system and using the same targets and optimization goals. We comparing the Homogeneity index (HI), Conformity index (CI) and sparing of the organs at risk for dose-volume histogram. Results: Whereas the HI, CI of TOMO was significantly better than the other, 3D-CRT was observed to have significantly poorer HI, CI. The percentage ipsilateral non-PTV breast volume that was delivered 50% of the prescribed dose was 3D-CRT (mean: 40.4%), TOMO (mean: 18.3%). The average ipsilateral lung volume percentage receiving 20% of the PD was 3D-CRT (mean: 4.8%), TOMO (mean: 14.2), concerning the average heart volume receiving 20% and 10% of the PD during treatment of left breast cancer 3D-CRT (mean: 1.6%, 3.0%), TOMO (mean: 9.7%, 26.3%) Conclusion: In summary, 3D-CRT and TOMO techniques were found to have acceptable PTV coverage in our study. However, in TOMO, high conformity to the PTV and effective breast tissue sparing was achieved at the expense of considerable dose exposure to the lung and heart.

• Investigative Magnetic Resonance Imaging
• /
• v.21 no.2
• /
• pp.65-70
• /
• 2017

Purpose: To optimize the saturation time and maximizing the pH-weighted difference between the normal and ischemic brain regions, on 3-tesla amide proton transfer (APT) imaging using an in vivo rat model. Materials and Methods: Three male Wistar rats underwent middle cerebral artery occlusion, and were examined in a 3-tesla magnetic resonance imaging (MRI) scanner. APT imaging acquisition was performed with 3-dimensional turbo spin-echo imaging, using a 32-channel head coil and 2-channel parallel radiofrequency transmission. An off-resonance radiofrequency pulse was applied with a Sinc-Gauss pulse at a $B_{1,rms}$ amplitude of $1.2{\mu}T$ using a 2-channel parallel transmission. Saturation times of 3, 4, or 5 s were tested. The APT effect was quantified using the magnetization-transfer-ratio asymmetry at 3.5 ppm with respect to the water resonance (APT-weighted signal), and compared with the normal and ischemic regions. The result was then applied to an acute stroke patient to evaluate feasibility. Results: Visual detection of ischemic regions was achieved with the 3-, 4-, and 5-s protocols. Among the different saturation times at $1.2{\mu}T$ power, 4 s showed the maximum difference between the ischemic and normal regions (-0.95%, P = 0.029). The APTw signal difference for 3 and 5 s was -0.9% and -0.7%, respectively. The 4-s saturation time protocol also successfully depicted the pH-weighted differences in an acute stroke patient. Conclusion: For 3-tesla turbo spin-echo APT imaging, the maximal pH-weighted difference achieved when using the $1.2{\mu}T$ power, was with the 4 s saturation time. This protocol will be helpful to depict pH-weighted difference in stroke patients in clinical settings.