• Korean Journal of Radiology
• /
• v.21 no.1
• /
• pp.15-24
• /
• 2020

Mammography is the primary screening method for breast cancers. However, the sensitivity of mammographic screening is lower for dense breasts, which are an independent risk factor for breast cancers. Automated breast ultrasound (ABUS) is used as an adjunct to mammography for screening breast cancers in asymptomatic women with dense breasts. It is an effective screening modality with diagnostic accuracy comparable to that of handheld ultrasound (HHUS). Radiologists should be familiar with the unique display mode, imaging features, and artifacts in ABUS, which differ from those in HHUS. The purpose of this study was to provide a comprehensive review of the clinical significance of dense breasts and ABUS screening, describe the unique features of ABUS, and introduce the method of use and interpretation of ABUS.

• Journal of Astronomy and Space Sciences
• /
• v.37 no.2
• /
• pp.95-104
• /
• 2020

Properties of plasmas that constitute the plasma sheet in the near-Earth magnetotail vary according to the solar wind conditions and location in the tail. In this case study, we present multi-spacecraft observations by Cluster that show a transition of plasma sheet from cold, dense to hot, tenuous state. The transition was associated with the passage of a spatial boundary that separates the plasma sheet into two regions with cold, dense and hot, tenuous plasmas. Ion phase space distributions show that the cold, dense ions have a Kappa distribution while the hot, tenuous ions have a Maxwellian distribution, implying that they have different origins or are produced by different thermalization processes. The transition boundary separated the plasma sheet in the dawn-dusk direction, and slowly moved toward the dawn flank. The hot, tenuous plasmas filled the central region while the cold, dense plasmas filled the outer region. The hot, tenuous plasmas were moving toward the Earth, pushing the cold, dense plasmas toward the flank. Different types of dynamical processes can be generated in each region, which can affect the development of geomagnetic activities.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.447-450
• /
• 2019

본 논문은 한국어 문장의 감성 분류를 위해 문장의 형태소, 음절, 자소를 입력으로 하는 합성곱층과 DenseNet 을 적용한 Text Multi-channel DenseNet 모델을 제안한다. 맞춤법 오류, 음소나 음절의 축약과 탈락, 은어나 비속어의 남용, 의태어 사용 등 문법적 규칙에 어긋나는 다양한 표현으로 인해 단어 기반 CNN 으로 추출 할 수 없는 특징들을 음절이나 자소에서 추출 할 수 있다. 한국어 감성분석에 형태소 기반 CNN 이 많이 쓰이고 있으나, 본 논문에서 제안한 Text Multi-channel DenseNet 모델은 형태소, 음절, 자소를 동시에 고려하고, DenseNet 에 정보를 밀집 전달하여 문장의 감성 분류의 정확도를 개선하였다. 네이버 영화 리뷰 데이터를 대상으로 실험한 결과 제안 모델은 85.96%의 정확도를 보여 Multi-channel CNN 에 비해 1.45% 더 정확하게 문장의 감성을 분류하였다.

• Journal of Korea Multimedia Society
• /
• v.23 no.12
• /
• pp.1540-1551
• /
• 2020

Aiming at the problem that the existing human behavior recognition algorithm cannot fully utilize the multi-level spatio-temporal information of the network, a human behavior recognition algorithm based on a dense three-dimensional residual network is proposed. First, the proposed algorithm uses a dense block of three-dimensional residuals as the basic module of the network. The module extracts the hierarchical features of human behavior through densely connected convolutional layers; Secondly, the local feature aggregation adaptive method is used to learn the local dense features of human behavior; Then, the residual connection module is applied to promote the flow of feature information and reduced the difficulty of training; Finally, the multi-layer local feature extraction of the network is realized by cascading multiple three-dimensional residual dense blocks, and use the global feature aggregation adaptive method to learn the features of all network layers to realize human behavior recognition. A large number of experimental results on benchmark datasets KTH show that the recognition rate (top-l accuracy) of the proposed algorithm reaches 93.52%. Compared with the three-dimensional convolutional neural network (C3D) algorithm, it has improved by 3.93 percentage points. The proposed algorithm framework has good robustness and transfer learning ability, and can effectively handle a variety of video behavior recognition tasks.

• Journal of The Korean Astronomical Society
• /
• v.24 no.2
• /
• pp.217-271
• /
• 1991

We have observed dense core around young stellar objects, DR21, S140, Orion-KL, and L1551 using four millimeter-wave transitions of $HC_3N\;J$=4-3, J=5-4, J=10-9, and J=12-11. The spatial distribution of $HC_3N$ emission closely resembles the morphology of the previous CS observations that trace high density gas. These observations reveal the existence of $HC_3N$ dense cores around central IR source, elliptical in shape and almost perpendicular to the CO bipolar outflow axis. Small differences can be explained by that $HC_3N$ molecular line is more optically thin and is seen to be more detailed structure in the neighborhood of central IR sources. In S140 and Orion-KL, massive(${\sim}10\;M_{\odot}$), slowly rotating dense cores lie near at the central IR sources of bipolar outflows. The velocity channel maps of DR21 show that the bipolar outflow gas may have a correlation with the dense core of DR21. We analyzed intensities of the four lines to derive physical conditions in dense core from two methods, LTE and LVG. The column density of $HC_3N$, $N(HC_3N)$, between LTE and LVG calculations agree well with each other. The abundances of $HC_3N$ in each observing source have been estimated using the average values of $n(H_2)$ and $N(HC_3N)$ and assuming the size of dense core. The fractional $HC_3N$ abundances in massive dense cores of DR21, S140, and Orion-KL have a range of $(2-7){\times}10^{-10}$, while that of low mass dense core, L1551, has one order of magnitude greater value of $2{\times}10^{-9}$. This should be considered good agreement with the result by Morris et al.(1976). It may be considered that dense cores of DR21, S140, and Orion-KL may have almost same stage of chemical evolution, and their abundances have a small values relative to that of L1551. The column density $N(HC_3N)$ decreases with increasing distance from the densest part of the cloud, the central infrared source, and have the relation of $N(HC_3N){\varpropto}R^{\alpha}$, where a has a range of 0.65 to 0.89. The values of $n(H_2)$ are not varied with increasing distance from the dense core, and have almost same values. Therefore, it is considered that the dense cores in these regions probably consist of dense clumps in diffuse molecular gas medium, and $n(H_2)$ of each clump is ${\sim}10^5\;cm^{-3}$. Levels in the $T_{ex}$ increases with $n(H_2)$. It is considered that the $HC_3N$ dense cores are not completely thermalized. We examine the relationships between the luminosity of central infrared sources versus mass of the dense cores, and the luminosity of central infrared sources versus molecular hydrogen column density. Luminosities of the central IR sources show good correlation with mass and hydrogen column density of the dense core. Same has been found from CS observations. However, mass and size derived from $HC_3N$ observations are one order of magnitude smaller than those from CS. It can be interpreted that we see more central part of the cloud cores in $NC_3N$ lines than CS lines.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.44 no.1
• /
• pp.20-25
• /
• 1999

This experiment was conducted to elucidate the relation-ship between vertical distribution of rice roots and yield traits under field conditions. Eight IRRI's new plant type rices (NPTRs) were tested in a volcanic ash soil paddy field under dense (IO 10 cm) and common (20 20 cm) planting densities. These lines were evaluated to have more spikelet numbers per panicle (SNP), lower filled grain rate (FGR), and lower rough grain weight per hill (RGWH). In dense planting, rough grain weight per stem (RGWS) was increased due to heavier culm and leaf dry weight (CLDW), and both RGWS and CLDW were related with the percentage of root distribution (%RWI) in the 10~30 cm soil layer, while in common planting, RGWS was not closely related with CLDW. SNP was highly related with root dry weight (RDW) in the 0~10cm soil layer. FGR was mainly affected by ROW in the 10~30 cm soil layer under both planting densities. RGWS was positively correlated with top dry weight (TDW) and harvest index (HI), and TDW was positively correlated with RWI under common planting or %RWI under dense planting, and HI was positively correlated with RWI in the 10~30 cm soil layer only under dense planting. RGWS was closely related with root weight index by dry weight (RWI) in the 10~30 cm soil layer and %RWI in the 0~30 cm or 10~30 cm soil layer under dense planting, and with only RWI in the 10~30 cm soil layer under common planting. But RGWH showed the close positive relationship with RDW and RWI in the 10~30 cm soil layer under dense planting, while under common planting, it showed the close positive relationship with RWI and %RWI in the 10~30 cm soil layer or %RWI in the 0~30 cm soil layer. The deeper root system in rice, especially under dense planting, is important for high yield of NPTRs focusing on the increment of top mass production and harvest index.

• Applied Microscopy
• /
• v.22 no.1
• /
• pp.1-14
• /
• 1992

Parenchyma of the cat pineal body consisted of pinealocytes and glial cells. The pinealocyte, predominant cell type, was characterized by having large mitochondria with pale matrix, abundant polyribosomes, moderately-developed Golgi apparatus, centrioles and occasional cilia. The pinealocyte had one thick and long cytoplasmic process at the one pole of the cell, and slender and shorter processes at the other pole, and in addition occasional short processes from the cell body. These processes contained longitudinally arranged microtubules, and a few mitochondria. Thick processes teminated as bulgings either in the intercellular process-rich area, or in the perivascular border which was formed by glial cell processes. These endings of pinealocyte processes had many small vesicles, mitochondria, and occasional dense bodies. Glial cells with abundant filaments of intermediate type and clear cytoplasmic matrix were fibrous astrocyte. Perikarya of the astrocytes had small and dense mitochondria, moderately developed Golgi apparatus, dense bodies and variable amount of intermediate filaments. Glial cell processes run through the intercellular spaces among the pinealocyte processes. Glial cell of protoplasmic type had no or a few filaments, but it had well-organized rough endoplasmic reticulum, dense mitochondria, well developed Golgi apparatus and many dense granules. Intercellular canaliculi formed by adjacent pinealocytes and glial cell processes were often noted. Within the parenchyma, sympathetic and parasympathetic axons and their endings were noted. These endings were present mostly in the intercellular spaces without having membrane specialization, however, in rare instances, ending with small clear and dense cored vesicles, and large dense cored vesicles formed specialized synapse with a pinealocyte process. Within the perivascular spaces nerve fibers and endings, Schwann cells and pericyte were noted. In rare case pinealocyte process penetrated into the perivascular space through the interuptions of glial border. These results suggest that pinealocyte of the cat has less significance in secretory function and is rather neural type of cell.

• Corrosion Science and Technology
• /
• v.9 no.5
• /
• pp.196-200
• /
• 2010

Recently porous metal has been used as supporting metal in planar type SOFC. In order to search optimum alloys for porous metal support and estimate the stability of metal-supported SOFC at high temperature, it is necessary to investigate the oxidation behaviors of porous material for metal support in comparison with dense material. Oxidation tests of porous and dense stainless steels were conducted at $600^{\circ}C$ and $800^{\circ}C$. Since the specific surface area of porous material is much larger than that of dense material, surface area should be considered in order to compare the oxidation rate of porous stainless steel with that of dense stainless steel. The specific surface area of porous body was measured using image analyzer. The weight gain of porous stainless steel was much greater than those of dense stainless steels due to its larger specific surface area. considering the specific surface area, the oxidation rate of porous stainless steel is likely to be the same as that of dense stainless steel with the same surface area. The change in chromium content in stainless steel during oxidation was also investigated. The experimental result in chromium content in stainless steel during oxidation corresponded with the calculated value. While the change in chromium content in dense stainless steel during oxidation is negligible, chromium content in porous stainless steel rapidly decreases with oxidation time due to its large specific surface area. The significant decrease in chromium content in porous stainless steel during oxidation may affect the oxidation resistance of porous stainless steel support and long term stability of metal-supported SOFC.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.12
• /
• pp.561-568
• /
• 2021

In recent years, autonomous driving technologies have become a high-value-added technology that attracts attention in the fields of science and industry. For smooth Self-driving, it is necessary to accurately detect an object and estimate its movement speed in real time. CNN-based deep learning algorithms and conventional dense optical flows have a large consumption time, making it difficult to detect objects and estimate its movement speed in real time. In this paper, using a single camera image, fast object detection was performed using the YOLOv5 algorithm, a deep learning algorithm, and fast estimation of the speed of the object was performed by using a local dense optical flow modified from the existing dense optical flow based on the detected object. Based on this algorithm, we present a system that can predict the collision time and probability, and through this system, we intend to contribute to prevent tram accidents.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.150-153
• /
• 2022

This paper reports a technique that automatically extracts object shapes through Dense-Net, and subsequently, detects the objects using Mask R-CNN in a manufacturing site, in which workers and objects are mixed. It is based on the customized factory dataset by targeting workers, machines, tools, control boxes, and products as the objects. Mask R-CNN supports multi-object recognition as a well-known object recognition method, while Dense-Net effectively extracts a feature from multiple and overlapping objects. After immediate implementation using the two technologies, the object is naturally extracted from a still image of the manufacturing site to describe image. Afterwards, the result is planned to be used to detect workers' abnormal behavior by adding a label on the objects.