• BMB Reports
• /
• 제54권2호
• /
• pp.98-105
• /
• 2021

The clustered regularly interspaced short palindromic repeats (CRISPR) system is a family of DNA sequences originally discovered as a type of acquired immunity in prokaryotes such as bacteria and archaea. In many CRISPR systems, the functional ribonucleoproteins (RNPs) are composed of CRISPR protein and guide RNAs. They selectively bind and cleave specific target DNAs or RNAs, based on sequences complementary to the guide RNA. The specific targeted cleavage of the nucleic acids by CRISPR has been broadly utilized in genome editing methods. In the process of genome editing of eukaryotic cells, CRISPR-mediated DNA double-strand breaks (DSB) at specific genomic loci activate the endogenous DNA repair systems and induce mutations at the target sites with high efficiencies. Two of the major endogenous DNA repair machineries are non-homologous end joining (NHEJ) and homology-directed repair (HDR). In case of DSB, the two repair pathways operate in competition, resulting in several possible outcomes including deletions, insertions, and substitutions. Due to the inherent stochasticity of DSB-based genome editing methods, it was difficult to achieve defined single-base changes without unanticipated random mutation patterns. In order to overcome the heterogeneity in DSB-mediated genome editing, novel methods have been developed to incorporate precise single-base level changes without inducing DSB. The approaches utilized catalytically compromised CRISPR in conjunction with base-modifying enzymes and DNA polymerases, to accomplish highly efficient and precise genome editing of single and multiple bases. In this review, we introduce some of the advances in single-base level CRISPR genome editing methods and their applications.

• Radiation Oncology Journal
• /
• 제34권4호
• /
• pp.313-321
• /
• 2016

Purpose: Total scalp irradiation (TSI) is a rare but challenging indication. We previously reported that non-coplanar intensity-modulated radiotherapy (IMRT) was superior to coplanar IMRT in organ-at-risk (OAR) protection and target dose distribution. This consecutive treatment planning study compared IMRT with volumetric-modulated arc therapy (VMAT). Materials and Methods: A retrospective treatment plan databank search was performed and 5 patient cases were randomly selected. Cranial imaging was restored from the initial planning computed tomography (CT) and target volumes and OAR were redelineated. For each patients, three treatment plans were calculated (coplanar/non-coplanar IMRT, VMAT; prescribed dose 50 Gy, single dose 2 Gy). Conformity, homogeneity and dose volume histograms were used for plan. Results: VMAT featured the lowest monitor units and the sharpest dose gradient (1.6 Gy/mm). Planning target volume (PTV) coverage and homogeneity was better in VMAT (coverage, 0.95; homogeneity index [HI], 0.118) compared to IMRT (coverage, 0.94; HI, 0.119) but coplanar IMRT produced the most conformal plans (conformity index [CI], 0.43). Minimum PTV dose range was 66.8%-88.4% in coplanar, 77.5%-88.2% in non-coplanar IMRT and 82.8%-90.3% in VMAT. Mean dose to the brain, brain stem, optic system (maximum dose) and lenses were 18.6, 13.2, 9.1, and 5.2 Gy for VMAT, 21.9, 13.4, 14.5, and 6.3 Gy for non-coplanar and 22.8, 16.5, 11.5, and 5.9 Gy for coplanar IMRT. Maximum optic chiasm dose was 7.7, 8.4, and 11.1 Gy (non-coplanar IMRT, VMAT, and coplanar IMRT). Conclusion: Target coverage, homogeneity and OAR protection, was slightly superior in VMAT plans which also produced the sharpest dose gradient towards healthy tissue.

• 한국산학기술학회논문지
• /
• 제16권12호
• /
• pp.8908-8914
• /
• 2015

인터넷이 지속적으로 성장과 발전을 거듭해가고 있는 이면에는 이를 악용하기 위한 다양한 인터넷 공격들이 발생하고 있다. 초기 인터넷 환경에서는 공격자가 역량과시 및 취미 등으로 인터넷 환경을 악용한 공격이 존재하였지만, 금전적인 이득을 목적으로 각종범죄와 연관된 체계적으로 복잡한 공격들이 발생하고 있다. 최근 들어서 바이러스나 윔과 같은 구조가 단순한 소스 멀티타깃(one source multi-target)의 형태가 존재하였지만, 멀티소스 싱글타깃(multi-source single target)의 형태를 갖는 APT(Advanced Persistent Threat, 지속적인 지능형 공격)으로 사용자들로 하여금 막대한 피해를 입히고 있다. 그러므로 본 논문에서는 Agent 및 관리 시스템은 악성코드 감염을 사전에 예방하는 기능을 고도화하여 사용자의 부정행위를 통한 자료유출을 감시할 수 방지 시스템을 설계 및 구현하였다. 성능평가에서는 감사데이터 생성 여부, 무결성 침해 발생 시 탐지 여부, 정상트래픽 오탐 여부, 프로세스 탐지 및 차단 기능 설정, Agent 정책 적용 가능여부에 대해서 기능을 분석하였다.

• 대한전자공학회논문지SD
• /
• 제44권1호
• /
• pp.68-75
• /
• 2007

NFC의 능동동작 모드, 수동동작 모드 및 RFID 동작 모드에 필요한 13.56MHz 트랜스미터와 리시버 및 RFID 태그 동작을 모두 지원하는 단일 칩 NFC 트랜시버를 설계 및 제작하고 동작을 검증하였다. 제안된 NFC 트랜시버는 외부전원 공급이 없어도 RFID 태그가 동작할 수 있도록 이니시에이터와 타겟의 2중 안테나 구조를 가지고 있다. 타겟 안테나는 이니시에이터 안테나의 접지 차폐층을 사용함으로써 이중 안테나의 유효면적이 단일 안테나에 비교해서 동일한 면적을 갖도록 안테나 구조를 제안하였고, 안테나의 선택 동작에 필요한 회로를 제안하였다. 제안된 NFC 단일 칩 트랜시버의 아날로그 전단부 회로는 능동모드와 RFID 리더를 위한 Reader/writer 블록의 트랜스미터와 리시버 회로부, 수동 모드와 태그 모드를 위한 태그 회로부로 구성된다. 태그 회로부는 정류기 및 부하 변조를 위한 수동소자가 포함되어 있으며, 정류기에서 생성되는 전압을 사용하여 외부 전원 없이도 태그 동작이 가능하도록 설계하였다. 제안된 트랜시버는 UART 직렬 인터페이스 회로를 통하여 호스트와 최대 212Kbps로 통신할 수 있다. 제안된 회로는 매그나칩의 0.35um 2-Poly 4-Metal CMOS공정으로 제작되었고, 칩의 유효면적은 $2200um{\times}360um$이다.

• 한국표면공학회지
• /
• 제27권2호
• /
• pp.65-73
• /
• 1994

Effects of deposition parameter on the ionic conductivity and structural change of the Lithium borosili-cate solid electrolyte films, prepared by rf sputtering using 7$LI_2O-3B_2O_3-1SiO_2$ single phase target and also a mosaic target enriched with $LI_2O$, were analyzed by measuring AC impedance and IR absorption spectra for the films. Thed solid electrolyte film deposited from the single phase target exhibited very low ionic conductivi-ty of $10^{-10}{\Omega}^{-1}cm{-1}$ at room temperature, a result of low $Li^+$ ion content(7.52 at%) in the film. The $Li^+$ con-ductivity for the films deposited from the mosaic target, however, significantly increased to $10^{-7}{\Omega}^{-1}cm{-1}$ due to both an increased $Li^+$content (14.75 at %) and a structural change of the films. The increased ionic conduc-tivity of the film appears to be associated with an easiness of ionic mobility by structural change of glassy film from a some close packed network structure to a open one. These structural changes of film were found to be closely related to the increase in the peak intensity at~$960cm^{-1}$ of IR absorption spectra for the glassy films. With increasing either argon pressure from 3 to 21 mtorr or rf power from 2 to 3 W/$cm^2$, the $Li^+$ conduc-tivity for the films significantly increased to an order of $10^{-6}{\Omega}^{-1}cm{-1}$ due to an increase in openness of film structure, as confirmed by both an increase in the IR absorption peak intensity at ~$960cm^{-1}$ and a resultant reduction of activation energy for mobility of $Li^+$ ion.

• Radiation Oncology Journal
• /
• 제9권2호
• /
• pp.351-359
• /
• 1991

LINAC 뇌정위적 방사선 수술은 주로 multiple noncoplanar arc를 이용한 single isocenter를 이용하고 있다. 이러한 방법에 의하면 구형선량 분포를 얻게 되는데, 뇌동정맥 기형이나 뇌종양의 경우 구형이 아닌 임의의 형태를 가진 경우도 많다. 본 논문에서는 임의의 병소형태에 대하여 multiple isocenters를 이용하여 병소모양과 같은 형태의 선량분포를 얻는 방법에 대하여 논하고져 한다. 적당한 조사변수들을 처음에 잘 선정하는 것은, 빠른 시간내에 최적선량 계획을 수행하는데 중요하다. 긴 병소모양에 대하여 같은 형태의 선량분포를 얻기 위한 isocenter 간격 및 조사면에 대한 guideline이 만들어졌다. 특히, 3차원 선량분포를 이용한 multiple isocenters의 응용에 대하여 논하여진다.

• 한국멀티미디어학회논문지
• /
• 제25권10호
• /
• pp.1375-1385
• /
• 2022

In this paper, we propose a semi-supervised domain adaptation solution to deal with practical face recognition (FR) scenarios where a single face image for each target identity (to be recognized) is only available in the training phase. Main goal of the proposed method is to reduce the discrepancy between the target and the source domain face images, which ultimately improves FR performances. The proposed method is based on the Domain Adatation network (DAN) using an MMD loss function to reduce the discrepancy between domains. In order to train more effectively, we develop a novel loss function learning strategy in which MMD loss and cross-entropy loss functions are adopted by using different weights according to the progress of each epoch during the learning. The proposed weight adoptation focuses on the training of the source domain in the initial learning phase to learn facial feature information such as eyes, nose, and mouth. After the initial learning is completed, the resulting feature information is used to training a deep network using the target domain images. To evaluate the effectiveness of the proposed method, FR performances were evaluated with pretrained model trained only with CASIA-webface (source images) and fine-tuned model trained only with FERET's gallery (target images) under the same FR scenarios. The experimental results showed that the proposed semi-supervised domain adaptation can be improved by 24.78% compared to the pre-trained model and 28.42% compared to the fine-tuned model. In addition, the proposed method outperformed other state-of-the-arts domain adaptation approaches by 9.41%.

• 융합신호처리학회논문지
• /
• 제11권3호
• /
• pp.203-208
• /
• 2010

본 논문에서는 소나 시스템을 통해 획득된 수중 음향 신호를 디지털 영상의 형태로 변환한다. 그리고 이러한 형태의 영상에 대해 영상 처리 기법을 도입하여 표적 후보를 탐지하고, 이들 영역에 대해 정보를 강화하는 알고리즘을 제안한다. 수중 표적의 탐지 과정은 우선 수중음향신호 영상에서 불규칙한 형태로 분포하고 있는 배경 잡음을 추정하여 재구성한 뒤, 원 영상에서 배경 영상을 제거하여 초기 표적 후보군을 획득한다. 또한 도플러 신호 정보를 가공하여 가중치 맵을 생성하고, 배경잡음이 제거된 영상에 대해 가중치 맵을 이용한 필터링 과정을 수행함으로써 표적 후보에 대한 정보를 보다 정확히 확보하고, 단일프레임에서의 표적 후보 정보를 강화한다. 본 논문에서는 시뮬레이션으로 획득된 수중음향신호에 대해 제안된 알고리즘을 적용하여, 불규칙적으로 발생하게 되는 잡음이 대부분 제거됨을 확인하였고, 필터링 및 표적 탐지 과정을 통해 수중음향신호 영상에서 표적이 더욱 명확히 표시됨을 확인하였다.

• 한국기계가공학회지
• /
• 제12권4호
• /
• pp.29-33
• /
• 2013

Here we report a partially aminated micropattern via direct functionalization and examine eleven different solution-phase probe DNAs hybridizing with the same target DNA on both hydrogen and oxygen terminated diamond. The hybridization intensities determined by epifluorescence microscopy were compared and are influenced strongly by the negatively charged surface and mismatched position of its sequence with immobilized probe DNA.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1991년도 한국자동제어학술회의논문집(국내학술편); KOEX, Seoul; 22-24 Oct. 1991
• /
• pp.687-692
• /
• 1991

To reduce the complexity in constructing a fuzzy logic controller of multivariable systems, three major methods are presented. One is the method of constructing single-input-single-output fuzzy logic controllers after decoupling the target system. Another is the method of using fuzzy relation matrices which indicate the relation between each input and output. The other is the method of using the hierarchically classified inputs which dominantly influence one output than other inputs. Using the last two methods, simulation results of fuzzy logic controller implemented on 160MW boiler-turbine plant model are also shown.