http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
CFG, 라이브러리 정보를 이용한 권한 기반 안드로이드 악성코드 탐지 기술의 성능 향상
박건우,Tamer AbuHmed,민대홍,양대헌,이경희 한국차세대컴퓨팅학회 2019 한국차세대컴퓨팅학회 논문지 Vol.15 No.6
In this paper, we propose a static Android malware detection technique to improve the detection performance and reliability. Based on the permission feature which is heavily used in the previous works, we additionally use the library dependency and control flow graph (CFG) as features for improving our detection performance. Library dependency and CFG-based features are efficient to detect Android malware, which is obfuscated using renaming technique because these are extracted by structural analysis. By combining these three features, we propose a novel malware detection model using bidirectional long short-term memory. As results, we achieved 99.62% overall detection rate. Our model is highly reliable: where the precision, recall and F1 scores are 100%, 99.26% and 99.62%, respectively. 본 논문에서는 안드로이드 악성코드의 증가 추세에 대응하여 향상된 성능의 정적 악성코드 탐지 기법을 고안하였다. 기존의 어플리케이션의 권한을 특징(feature)으로 사용하는 악성코드 탐지 기법에 라이브러리 사용 정보와 control flow graph (CFG)의 속성을 특징으로 추가하여 성능을 향상시켰다. 또한, 라이브러리와 CFG는 구조 분석을 통해 특징을 추출하므로 리네이밍(renaming) 난독화에 대하여 독립적이라는 특징이 있어 난독화에 취약한 권한 사용 탐지 기법을 보완하는 추가적인 이점을 가진다. 어플리케이션으로부터 추출한 세 가지 특징을 기반으로 양방향 장단기 기억 네트워크(bidirectional long short-term memory)를 이용한 악성코드 탐지 모델을 제안하였다. 세 가지 특징을 모두 사용한 악성코드 분류 모델을 안드로이드 악성코드와 일반 어플리케이션을 합친 데이터에 적용하였을 때 정확도 99.62%, 정밀도 100%, 재현율 99.26%, F1 99.62%로 높은 성능과 신뢰도를 보였다.
이영진(Y . J . Lee),황선욱(S . W . Hwang),김상권(S . K . Kim),박윤서(Y . S . Park),전재근(J . K . Chun) 한국응용생명화학회 1995 Applied Biological Chemistry (Appl Biol Chem) Vol.38 No.5
In order to investigate the textural changes of cooked rice in a home electrical rice cooker during warm keeping period, rice eras cooked and stored at 73℃ for 42 hours. The time course changes of hardness and adhesiveness of cooked rice were measured with the specimen prepared as a cylindrical rice block(H/D=15 ㎜/20 ㎜). At the initial stage of warm keeping period the hardness was 1.61 ㎏ and it was increased steadily to 1.95 ㎏ up to 12 hours. And it was rapidly dropped to 1.57 ㎏ within 3 hours, and thereafter gradually decreased to 1.44 ㎏. The adhesiveness was steadily declined with warm keeping period from 0.51 to 0.323 ㎏·sec. The cohesiveness and chewiness were also changed in the similar pattern to that of hardness while the springiness remained unchanged throughout the whole warm keeping period.
폴리트리메틸렌텔레프탈레이트 / 폴리비스페놀카보네이트 용융 블렌드 (1)
나상권,김정규,홍완해 한국공업화학회 2000 응용화학 Vol.4 No.2
Melt blends of poly(trimethylene terephthalate), PTT and Poly(Bisphenol A carbonate), PC with various compositions were prepared at 275℃ using 40rpm for 60min. Thermal properties of PTT/PC blends have been studied by differential scanning calorimetry(DSC) and dynamic mechanical thermal analysis(DMTA). The glass transition temperatures of blends increased as the amount of PC increased. Each blend showed a single peak of tan δ in the glass transition region, suggesting that PTT and PC might be compatible.