• The Korean Journal of Food And Nutrition
• /
• v.37 no.4
• /
• pp.197-210
• /
• 2024

This study aims to develop a deep learning model to monitor rice serving amounts in institutional foodservice, enhancing personalized nutrition management. The goal is to identify the best convolutional neural network (CNN) for detecting rice quantities on serving trays, addressing balanced dietary intake challenges. Both a vanilla CNN and 12 pre-trained CNNs were tested, using features extracted from images of varying rice quantities on white trays. Configurations included optimizers, image generation, dropout, feature extraction, and fine-tuning, with top-1 validation accuracy as the evaluation metric. The vanilla CNN achieved 60% top-1 validation accuracy, while pre-trained CNNs significantly improved performance, reaching up to 90% accuracy. MobileNetV2, suitable for mobile devices, achieved a minimum 76% accuracy. These results suggest the model can effectively monitor rice servings, with potential for improvement through ongoing data collection and training. This development represents a significant advancement in personalized nutrition management, with high validation accuracy indicating its potential utility in dietary management. Continuous improvement based on expanding datasets promises enhanced precision and reliability, contributing to better health outcomes.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.158-161
• /
• 2003

A cascade structured neural network called Sigma-Pi$_{t}$ Cascaded Hybrid Neural Network ($\sigma$$\pi$$_{t}$-CHNN) is Proposed. It is an extended version of the Sigma-Pi Cascaded extended Hybrid Neural Network ($\sigma$$\pi$-CHNN), where the classical multiplicative neuron ($\pi$-neuron) is replaced by the translated multiplicative ($\pi$$_{t}$-neuron) model. The learning algorithm of $\sigma$$\pi$$_{t}$-CHNN is composed of an evolutionary programming method, responsible for determining the network architecture, and of a Levenberg-Marquadt algorithm, responsible for tuning the weights of the network. The $\sigma$$\pi$$_{t}$-CHNN is evaluated in 2 pattern classification problems: the 2 spirals and the sonar problems. In the 2 spirals problem, $\sigma$$\pi$$_{t}$-CHNN can generate neural networks with 10％ less hidden neurons than that in previous neural models. In the sonar problem, $\sigma$$\pi$$_{t}$-CHNN can find the optimal solution for the problem i.e., a network with no hidden neurons. These results confirm the expanded information processing capabilities of $\sigma$$\pi$$_{t}$-CHNN, when compared to previous neural network models. network models.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.18 no.3
• /
• pp.345-355
• /
• 2008

This paper presents a design of a tuned liquid mass damper(TLMD) for controlling bi-directional response of high-rise building structure subjected to windload. The proposed damper behaves as a tuned mass damper(TMD) of which mass is regarded as the mass of a tuned liquid column damper(TLCD) and the case wall of the TLCD itself in one direction and the TLCD in the other direction. Because the proposed device has coupled design parameter along two orthogonal directions, it is very important to select designing components by optimal fine tuning. In the designing TLMD, for easy maintenance, the rubber-bearing with small springs was applied in TMD direction. In this study, the Songdo New City Tower 1A in Korea, which has been designed and constructed two TLCDs in order to control bi-directional response, was chosen as the model building structure. The results of rotation test proved the effectiveness of bi-directional behavior of TLMD.

• Tribology and Lubricants
• /
• v.32 no.2
• /
• pp.61-66
• /
• 2016

Total thickness variation (TTV), BOW, and surface roughness are essential characteristics for high quality sapphire substrates. Many researchers have attempted to increase removal rate by controlling the key process parameters like pressure and velocity owing to the high cost of consumables in sapphire chemical mechanical polishing (CMP). In case of the pressure approach, increased pressure owing to higher deviation of pressure over the wafer leads to significant degradation of the TTV. In this study, the authors focused on reducing TTV under the high-pressure conditions. When the production equipment polishes multiple wafers attached on a carrier, higher loads seem to be concentrated around the leading edge of the head; this occurs because of frictional force generated by the combination of table rotation and the height of the gimbal of the polishing head. We believe the skewed pressure distribution during polishing to be the main reason of within-wafer non-uniformity (WIWNU). The insertion of a hub ring between the polishing head and substrate carrier helped reduce the pressure deviation. Adjusting the location of the hub ring enables tuning of the pressure distribution. The results indicated that the position of the hub ring strongly affected the removal profile, which confirmed that the position of the hub ring changes the pressure distribution. Furthermore, we analyzed the deformation of the head via finite element method (FEM) to verify the pressure non-uniformity over the contact area Based on experiment and FEM results, we determined the optimal position of hub ring for achieving uniform polishing of the substrate.

• Journal of the Korean Wood Science and Technology
• /
• v.45 no.6
• /
• pp.797-808
• /
• 2017

Automatic wood species identification systems have enabled fast and accurate identification of wood species outside of specialized laboratories with well-trained experts on wood species identification. Conventional automatic wood species identification systems consist of two major parts: a feature extractor and a classifier. Feature extractors require hand-engineering to obtain optimal features to quantify the content of an image. A Convolutional Neural Network (CNN), which is one of the Deep Learning methods, trained for wood species can extract intrinsic feature representations and classify them correctly. It usually outperforms classifiers built on top of extracted features with a hand-tuning process. We developed an automatic wood species identification system utilizing CNN models such as LeNet, MiniVGGNet, and their variants. A smartphone camera was used for obtaining macroscopic images of rough sawn surfaces from cross sections of woods. Five Korean softwood species (cedar, cypress, Korean pine, Korean red pine, and larch) were under classification by the CNN models. The highest and most stable CNN model was LeNet3 that is two additional layers added to the original LeNet architecture. The accuracy of species identification by LeNet3 architecture for the five Korean softwood species was 99.3%. The result showed the automatic wood species identification system is sufficiently fast and accurate as well as small to be deployed to a mobile device such as a smartphone.

• The Journal of Information Technology
• /
• v.4 no.2
• /
• pp.117-127
• /
• 2001

A self-structured neural network algorithm that finds optimal fuzzy membership functions and nile base to fuzzy model is proposed and a fuzzy-neural network controller is designed to get more accurate position and velocity control of wheeled mobile robot. This procedure that is composed of three steps has its own unique process at each step. The elements of output term set are increased at first step and then the rule base Is varied according to increase of the elements. The adjusted controller is in competition with controller which doesn't include any increased elements. The adjusted controller will be removed if the control-law lost. Otherwise, the controller is replaced with the adjusted system. After finished regulation of output term set and rule base, searching for input membership functions is processed with constraints and fine tuning of output membership functions is done.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.35 no.5
• /
• pp.438-445
• /
• 2007

A Helmholtz resonator as a stabilization device to control high-frequency combustion instabilities in liquid rocket engine is adopted and its damping capacity is verified by linear acoustic analysis and atmospheric acoustic tests. To compare the results of acoustic attenuation effect in accordance with uni-resonator's geometry, quantitative analyses were made in the cases of various orifice diameters and lengths. Next, in the experiments to compare the results of acoustic attenuation effect by a difference in the number of resonators, damping capacity of harmful resonant frequency was improved by the increase of the number of resonators. On the other hand, attenuation efficiency of the frequency tended rather to lower due to over damping from the point of view of absorption coefficient. As the result, tuning the suitable geometry for the resonator to the resonant frequency is required for the control using the resonator. Also, the design of resonator's geometry and the choice of its number are important to put up the optimal efficiency in consideration of restriction of its volume.

• Nuclear Engineering and Technology
• /
• v.52 no.2
• /
• pp.279-286
• /
• 2020

In this article, we develop a reactive distillation (RD) column configuration for the production of hydrogen. This RD column is in the HI decomposition section of the sulphur - iodine (SI) thermochemical cycle, in which HI decomposition and H₂ separation take place simultaneously. The section plays a major role in high hydrogen production efficiency (that depends on reaction conversion and separation efficiency) of the SI cycle. In the column simulation, the rigorous thermodynamic phase equilibrium and reaction kinetic model are used. The tuning parameters involved in phase equilibrium model are dependent on interactive components and system temperature. For kinetic model, parameter values are adopted from the Aspen flowsheet simulator. Interestingly, there is no side reaction (e.g., solvation reaction, electrolyte decomposition and polyiodide formation) considered aiming to make the proposed model simple that leads to a challenging prediction. The process parameters are determined on the basis of optimal hydrogen production as reflux ratio = 0.87, total number of stages = 19 and feeding point at 8th stage. With this, the column operates at a reasonably low pressure (i.e., 8 bar) and produces hydrogen in the distillate with a desired composition (H₂ = 9.18 mol%, H₂O = 88.27 mol% and HI = 2.54 mol%). Finally, the results are compared with other model simulations. It is observed that the proposed scheme leads to consume a reasonably low energy requirement of 327 MJ/kmol of H₂.

• Proceedings of the KIEE Conference
• /
• 2008.07a
• /
• pp.1297-1300
• /
• 2008

A conventional linear accelerator system requires a flat-topped pulse with less than ${\pm}$ 0.5% ripple to meet the beam energy spread requirements and to improve pulse efficiency of RF systems. A pulse transformer is one of main determinants on the output pulse voltage shape. The pulse transformer was investigated and analyzed with the pulse response characteristics using a simplified equivalent circuit model. The damping factor ${\sigma}$ must be >0.86 to limit the overshoot to less than 0.5% during the flat-top phase. The low leakage inductance and distributed capacitance are often limiting factors to obtain a fast rise time. These parameters are largely controlled by the physical geometry and winding configuration of the transformer. A rise time can be improved by reducing the number of turns, but it produces larger pulse droop and requires a larger core size. By tradeoffs among these parameters, the high-voltage pulse transformer with a pulse width of 10 ${\mu}s$, a rise time of 0.84 ${\mu}s$, and a pulse droop of 2.9% has been designed and fabricated to drive a klystron which has an output voltage of 284 kV, 30-MW peak and 60-kW average RF output power. This paper describes design optimization of a high-voltage pulse transformer for high-power pulsed applications. The experimental results were analyzed and compared with the design. The design and optimal tuning parameter of the system was identified using the model simulation.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.22 no.9
• /
• pp.1243-1249
• /
• 2018

In this paper, we propose a scheme that tunes the backoff period for enhancing the system throughput with estimating the number of devices in IEEE 802.15.4 slotted carrier sense multiple access with collision avoidance (CSMA/CA) networks. Since each device does not sense the channel always in IEEE 802.15.4 slotted CSMA/CA networks, a personal area network (PAN) coordinator is used to estimate the number of active devices. The PAN coordinator broadcasts an optimal backoff period for the estimated number of devices through a beacon frame. In order to estimate the number of devices in run time, a simple moving average filter is utilized. We show the performance of our proposed scheme in terms of the estimated number of devices and the system throughput. The simulation results show that our proposed scheme can obtain higher system throughput than the IEEE 802.15.4 standard.