• Journal of IKEEE
• /
• v.26 no.4
• /
• pp.647-652
• /
• 2022

In this paper, we propose a prediction system for skin pore labeling based on a CNN(Convolution Neural Network) model, where a data set is constructed by processing skin images taken by users, and a pore feature image is generated by the proposed image processing algorithm. The skin image data set was labeled for pore characteristics based on the visual classification criteria of skin beauty experts. The proposed image processing algorithm was applied to generate pore feature images from skin images and to train a CNN model that predicts pore feature ratings. The prediction results with pore features by the proposed CNN model is similar to experts visual classification results, where less learning time and higher prediction results were obtained than the results by the comparison model (Resnet-50). In this paper, we describe the proposed image processing algorithm and CNN model, the results of the prediction system and future research plans.

• Proceedings of the Korea Association of Crystal Growth Conference
• /
• 1996.06a
• /
• pp.494-496
• /
• 1996

The purpose of this study is to investigate the effect of process conditions on pore distribution in porous silicon layer prepared by electrochemical reaction. Porous silicon layers formed on p-type silicon wafer show the network structure of fine porse whose diameters are less than 100${\AA}$. In n-type porous silicon, selective growth was found on the pore surface by wet etching process after PR patterning. And numerical method showed high current density on the pore tip. With this result we confirmed that pore formation has two steps. First step is the initial attack on the surface and second step is the directional growth on the pore tip.

• Journal of the Korea Concrete Institute
• /
• v.16 no.2 s.80
• /
• pp.277-282
• /
• 2004

In a hardened concrete, diffusion of oxygen, carbon dioxide, aggressive ions, and moisture from the environment to the concrete takes place through the pore network. It is well known that making dense cement matrix enhances the durability of concrete as well as all the characteristics including strength of concrete. In this paper,9 mix concretes with water to cementitious material ratio (40,45, and $50\%$) and replacement ratio of GGBFS (40 and $60\%$ of cement by weight) were studied on the micro-pore structure by mercury intrusion porosimetry and the accelerated chloride diffusion test by potential difference. From the results the average pore diameter and accelerated chloride diffusivity of concrete were ordered NPC > G4C > G6C. It is concluded that there is a good correlation between the average pore diameter and the chloride diffusivity, and the mineral admixtures has a filling effect, which increases the tortuosity of pore and makes large pores finer, on the pore structure of cement matrix due to the latent hydraulic reaction with hydrates of cement.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.25 no.2
• /
• pp.64-69
• /
• 2013

The structure of pore network of gas diffusion layers (GDLs) in PEMFCs plays a critical role in determining the transport phenomena of reaction gas as well as generated water. In addition, the interactive characteristics between water and surface of pore are no less important than the structural characteristics of pore network. In this study, porometric investigation is conducted for two kinds of GDL using method of standard porosimetry which enable to distinguish hydrophobic pores from hydrophilic pores of GDLs. The porosity of TGPH-120 decreases by 6% by adding 30 wt.% of PTFE, but the porosity of hydrophilic pores decreases by 12%. The relation of $p_c-S_{nw}$ varies with the addition of PTFE, especially at low $p_c$.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.16 no.4
• /
• pp.311-316
• /
• 2003

Effects of pores on the microwave properties in microwave dielectric ceramics were studied by a computer simulation. Scattering matrix S$\_$21/ obtained from the network analyzer was compared to the S$\_$21/ obtained from the simulation. From electric field distribution, the dominant resonant TE$\_$01$\delta$/ mode could be easily determined. The effects of the porosity and pore size inside the dielectrics on the microwave properties were investigated by the HFSS simulation. When the total pore volume remains constantly, the quality factor decreased as the pore size Increases. As the total pore volume of the dielectrics increased. quality factor decreased.

• Korean Journal of Materials Research
• /
• v.16 no.11
• /
• pp.715-718
• /
• 2006

Porous hydroxyapatite (HAp) scaffolds have been prepared by using the slurry including HAp and magnesia based on the replication of polymer sponge substrate. The influence of MgO content in slurry on the pore morphology and size, density, porosity, and mechanical strength of porous HAp scaffolds was investigated. The obtained scaffolds with average pore sizes ranging 150 to 300 mm had open, relatively uniform, and interconnected porous structure regardless of MgO content. As the MgO content increased, the pore network frame of scaffolds became to be relatively stronger, even though the pore size was not much changed. The compressive strength of the scaffolds increased rapidly with the increase of MgO content because of increasing the pore wall thickness and density of the scaffolds. As a result, the porosity, density, and compressive strength of the porous HAp scaffolds prepared by the sponge method were significantly affected by the addition of MgO.

• Journal of Industrial Technology
• /
• v.34
• /
• pp.39-43
• /
• 2014

This paper represents a highly porous MWCNT film electrode with interconnected open pores and demonstrated the possibility of using an MWCNT network film as the top electrode for polyimide capacitive humidity sensors. Polyimide humidity sensors with MWCNT electrodes exhibited about 6 times faster response than equivalent Cr electrode sensors. This result may be due to their percolated pore structures, which make water molecules accessible to all polyimide surfaces. The much faster response times of MWCNT electrode sensors is attributed to the percolated pore network, which allows more water molecules to be accessible to polyimide surfaces.

• Journal of the Korean Ceramic Society
• /
• v.40 no.12
• /
• pp.1159-1164
• /
• 2003

Two processing routes (i.e., the gel casting and polymer preform routes) using polymer beads were studied to fabricate porous ceramics with a cellular structure. The gel casting route, comprising the gel casting of a ceramic slurry mixed with polymer beads, was found to be inadequate to produce porous ceramic bodies with a interconnected pore structure, due to complete coating of the slurry on the polymer beads, which left just isolated pores in the final sintered bodies. The polymer preform route, involving the infiltration of a polymer beads preform with the ceramic slurry, successfully produced porous ceramics with a highly interconnected network of spherical pores. The pore size of 250-300 $\mu\textrm{m}$ was demonstrated and the porosity ranged from 82 to 86%. This process is advantageous to control the pore size because it is determined by the sizes of polymer beads used. Another feature is the avoidance of hollow skeleton, giving a high strength.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.5
• /
• pp.339-346
• /
• 2023

The uncertainties of microstructural features affect the properties of materials. Numerous pores that are randomly distributed in materials make it difficult to predict the properties of the materials. The distribution of pores in cementitious materials has a great influence on their mechanical properties. Existing studies focus on analyzing the statistical relationship between pore distribution and material responses, and the correlation between them is not yet fully determined. In this study, the mechanical response of cementitious materials is predicted through an image-based data approach using a convolutional neural network (CNN), and the correlation between pore distribution and material response is analyzed. The dataset for machine learning consists of high-resolution micro-CT images and the properties (tensile strength) of cementitious materials. The microstructures are characterized, and the mechanical properties are evaluated through 2D direct tension simulations using the phase-field fracture model. The attributes of input images are analyzed to identify the spot with the greatest influence on the prediction of material response through CNN. The correlation between pore distribution characteristics and material response is analyzed by comparing the active regions during the CNN process and the pore distribution.

• Journal of the Korean GEO-environmental Society
• /
• v.10 no.2
• /
• pp.61-68
• /
• 2009

This paper presents the numerical investigation for the hydraulic behavior of a fractured rock mass with a hydromechanical interaction which may be considered during the in-situ hydraulic injection test. These experiments consist in a series of flow meter injection tests for fractures existing along an open hole section installed in a borehole, and experimental results are applied for testing a numerical model developed to the analysis and prediction of such hydromechanical interactions. Field experimental results show that conductive fractures form a dynamic and interdependent network, that individual fractures cannot be adequately modeled as independent systems, that new fluid intaking zones generate when pore pressure exceeds the minimum principal stress magnitude in that borehole, and that pore pressures much larger than this minimum stress can be further supported by the circulated fractures. In this study, these characteristics are investigated numerically how to influence the morphology of the natural fracture network in a rock mass by using a discrete fracture ntework model.