• 정보보호학회논문지
• /
• 제30권3호
• /
• pp.337-347
• /
• 2020

RYPTO 2019에 발표된 Gohr의 연구결과는 딥러닝 기술이 암호분석에 활용될 수 있음을 보여주었다. 본 논문에서는 특정 구조를 가진 S-box를 딥러닝 기술이 식별할 수 있는지 실험한 결과를 제시한다. 이를 위해, 2가지 실험을 수행하였다. 첫 번째로는, 경량암호 설계에 주로 사용하는 Feistel 및 MISTY, SPN, multiplicative inverse 구조를 가진 S-box의 DDT 및 LAT로 학습 데이터를 구성하고 딥러닝 알고리즘으로 구조를 식별하는 실험을 수행하여 구조는 물론 라운드까지 식별할 수 있었다. 두 번째로는 Feistel 및 MISTY 구조가 특정 라운드까지 의사난수성을 보이는지에 대한 실험을 통해 이론적으로 제시된 라운드 수 보다 많은 라운드 수에서 random한 함수와 구분할 수 있음을 확인하였다. 일반적으로, 군사용 등 고도의 기밀성 유지를 위해 사용되는 암호들은 공격이나 해독을 근본적으로 차단하기 위해 설계정보를 공개하지 않는 것이 원칙이다. 본 논문에서 제시된 방법은 딥러닝 기술이 이처럼 공개되지 않은 설계정보를 분석하는 하나의 도구로 사용 가능하다는 것을 보여준다.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 1998년도 제16회 학술발표회 논문개요집
• /
• pp.96-96
• /
• 1998

• 한국방사성폐기물학회:학술대회논문집
• /
• 한국방사성폐기물학회 2018년도 추계학술논문요약집
• /
• pp.565-566
• /
• 2018

• The Journal of Korean Physical Therapy
• /
• 제27권1호
• /
• pp.50-54
• /
• 2015

Purpose: The purpose of this study is to investigate differences of cervical flexor muscle thickness (i.e., sternocleidomastoid muscle and deep cervical flexor muscles) depending on levels of pressure bio-feedback unit and eye directions during cranial-cervical flexor exercise in healthy subjects. Methods: A total of 30 subjects (12 males and 18 females) who had no medical history related to musculoskeletal and neurological disorders were enrolled in this study. They were instructed to perform cranial-cervical flexion exercise with adjustment of five different pressures (i.e., 22 mmHg, 24 mmHg, 26 mmHg, 28 mmHg, and 30 mmHg) using a pressure biofeedback unit, according to three different eye directions (i.e., $0^{\circ}$, $20^{\circ}C$, and $40^{\circ}C$). Muscle thickness of sternocleidomastoid muscle and deep cervical flexor muscles was measured according to pressure levels and eye directions using ultrasonography. Results: In results of muscle thickness in sternocleidomastoid muscle and deep cervical flexor muscles, the thickness of those muscles was gradually increased compared to the baseline pressure level (22 mmHg), as levels in the pressure biofeedback unit during cranial-cervical flexion exercise were increasing. In addition, at the same pressure levels, muscle thickness was increased depending on ascending eye direction. Conclusion: Our findings showed that muscle thickness of sternocleidomastoid muscle and deep cervical flexor muscles was generally increased during cranial-cervical flexion exercise, according to increase of eye directions and pressure levels. Therefore, we suggested that lower eye direction could induce more effective muscle activity than the upper eye direction in the same environment during cranial-cervical flexion exercise.

• 융합정보논문지
• /
• 제10권11호
• /
• pp.23-29
• /
• 2020

IoT 장치가 기하급수적으로 증가하면서 다중 안테나를 통해 IoT 정보를 송·수신하기 위한 MIMO 간섭 최소화 및 전송 용량 증대는 가장 큰 이슈로 남아있는 상황이다. 본 논문에서는 MIMO 기반의 IoT 통신 잡음을 최소화하기 위해서 딥러닝을 활용한 비밀키 차원 분배 메커니즘을 제안한다. 제안 메커니즘은 다중의 안테나를 통해 송·수신되는 IoT 정보를 딥러닝을 사용하여 일괄적으로 분산 처리함으로써 송·수신 과정 중에 발생하는 자원 손실을 최소화하고 있다. 또한, 제안 메커니즘은 AP들간의 직접적인 간섭이 없는 기지국의 다중 안테나 다중 스트림 전송을 통해 용량을 최대로 증대시킬 수 있도록 다차원 키 분배 처리 과정을 적용하였다. 또한, 제안 메커니즘은 다중 안테나 기술을 최대한 활용하기 위해서 IoT 정보의 주파수 채널 수에 따라 비밀키를 차원 분배하는 방식을 적용함으로써 IoT 정보수에 따른 비밀키 사용 빈도수를 딥러닝하여 IoT 정보를 서로 동기화하고 있다.

• Journal of Dental Anesthesia and Pain Medicine
• /
• 제16권1호
• /
• pp.49-53
• /
• 2016

Background: General anesthesia is frequently considered for pediatric patients, as they often find it difficult to cooperate and stay calm during administration of potentially painful treatments. Sedation can overcome these adversities; however, this is challenging while maintaining unobstructed airways. Methods: The study involved 11 pediatric dental patients treated with LMA under deep sedation with sevoflurane, from 2011 through 2015. LMA size, sevoflurane concentration, and the vital signs of patients were assessed through a chart review. Results: The age distribution of the patients ranged from 6 to 10 years old. A total of 3 patients underwent mesiodens extraction, while the remaining 8 underwent an surgically assisted orthodontic forced tooth eruption The average sedation period was approximately 45 minutes and the LMA size was $2\small{^1/_2}$. The sevoflurane concentration was maintained at 2% on average, and overall, the measurements of vital signs were within the normal range; the patients had an average blood pressure of 98/49 mmHg, breathing rate of 26 times/min, pulse frequency of 95 times/min, $SpO_2s$ level of 99 mmHg, and $ETCO_2$ level of 41.2 mmHg. Conclusions: Deep sedation with sevoflurane coupled with LMA may be applied successfully in pediatric patients who undergo mesiodens extraction or a surgically assisted orthodontic forced tooth eruption

• 터널과지하공간
• /
• 제29권2호
• /
• pp.75-88
• /
• 2019

우리나라는 1978년 4월 고리1호기를 시작으로 지금까지 총 24기의 원전을 가동하고 있으며 2기의 원전이 건설 중이다. 원자력 발전이 지속됨에 따라 원자력발전소에서 발생하는 방사성 폐기물의 양도 늘어나게 되어 이를 영구처분하기 위한 다양한 방법이 제안되어 왔다. 국내에서는 심층처분(DGD)을 중심으로 연구가 진행되어 왔으나 심부 시추공을 활용하는 심부시추공 처분(DBD) 역시 대안으로 고려할 필요가 있다. 본 논문에서는 기술 선진국의 선행 연구결과를 종합하여 심부시추공 처분에 요구되는 요소기술들을 소개하고 이를 국내에 적용하기 위한 적용성 평가를 수행하였다. 시추공 설계, 처분부지 등에 대한 개념적 연구를 수행하였으며 마지막으로 실제 처분을 위하여 향후 요구되는 기술적 과제에 대하여 정리하였다.

• 한국차세대컴퓨팅학회논문지
• /
• 제14권6호
• /
• pp.87-97
• /
• 2018

강우량은 매우 중요한 기상 정보이다. 일반적으로, 도로 수준과 같은 높은 공간 해상도의 강우량이 더 높은 가치를 가진다. 하지만, 도로 수준의 강우량을 측정하기 위해 충분한 수의 기상 관측 장비를 설치하는 것은 비용 관점에서 비효율적이다. 본 논문에서는 도로의 감시 카메라 영상으로부터 강우량을 인식하기 위해 심층 신경망을 활용하는 방법에 대해 제시한다. 해당 목표를 달성하기 위해, 본 논문에서는 교내 두 지역의 감시 카메라 영상과 강우량 데이터를 수집했으며, 새로운 심층 신경망 구조인 Temporal and Spatial Segment Networks(TSSN)를 제안한다. 본 논문에서 제시한 심층 신경망으로 강우량 인식을 수행한 결과, 프레임 RGB와 두 연속 프레임 RGB 차이를 입력으로 사용했을 때, 높은 성능으로 강우량 인식을 수행할 수 있었다. 또한, 기존의 심층 신경망 모델과 비교했을 때, 본 논문에서 제안하는 TSSN이 가장 높은 성능을 기록함을 확인할 수 있었다.

• Imaging Science in Dentistry
• /
• 제52권2호
• /
• pp.219-224
• /
• 2022

Purpose: This study aimed to evaluate the performance of transfer learning in a deep convolutional neural network for classifying implant fixtures. Materials and Methods: Periapical radiographs of implant fixtures obtained using the Superline (Dentium Co. Ltd., Seoul, Korea), TS III(Osstem Implant Co. Ltd., Seoul, Korea), and Bone Level Implant(Institut Straumann AG, Basel, Switzerland) systems were selected from patients who underwent dental implant treatment. All 355 implant fixtures comprised the total dataset and were annotated with the name of the system. The total dataset was split into a training dataset and a test dataset at a ratio of 8 to 2, respectively. YOLOv3 (You Only Look Once version 3, available at https://pjreddie.com/darknet/yolo/), a deep convolutional neural network that has been pretrained with a large image dataset of objects, was used to train the model to classify fixtures in periapical images, in a process called transfer learning. This network was trained with the training dataset for 100, 200, and 300 epochs. Using the test dataset, the performance of the network was evaluated in terms of sensitivity, specificity, and accuracy. Results: When YOLOv3 was trained for 200 epochs, the sensitivity, specificity, accuracy, and confidence score were the highest for all systems, with overall results of 94.4%, 97.9%, 96.7%, and 0.75, respectively. The network showed the best performance in classifying Bone Level Implant fixtures, with 100.0% sensitivity, specificity, and accuracy. Conclusion: Through transfer learning, high performance could be achieved with YOLOv3, even using a small amount of data.

• 인터넷정보학회논문지
• /
• 제25권2호
• /
• pp.11-19
• /
• 2024

This paper was studied abouta technology for detecting damage to temporary works equipment used in construction sites with explainable artificial intelligence (XAI). Temporary works equipment is mostly composed of steel or aluminum, and it is reused several times due to the characters of the materials in temporary works equipment. However, it sometimes causes accidents at construction sites by using low or decreased quality of temporary works equipment because the regulation and restriction of reuse in them is not strict. Currently, safety rules such as related government laws, standards, and regulations for quality control of temporary works equipment have not been established. Additionally, the inspection results were often different according to the inspector's level of training. To overcome these limitations, a method based with AI and image processing technology was developed. In addition, it was devised by applying explainableartificial intelligence (XAI) technology so that the inspector makes more exact decision with resultsin damage detect with image analysis by the XAI which is a developed AI model for analysis of temporary works equipment. In the experiments, temporary works equipment was photographed with a 4k-quality camera, and the learned artificial intelligence model was trained with 610 labelingdata, and the accuracy was tested by analyzing the image recording data of temporary works equipment. As a result, the accuracy of damage detect by the XAI was 95.0% for the training dataset, 92.0% for the validation dataset, and 90.0% for the test dataset. This was shown aboutthe reliability of the performance of the developed artificial intelligence. It was verified for usability of explainable artificial intelligence to detect damage in temporary works equipment by the experiments. However, to improve the level of commercial software, the XAI need to be trained more by real data set and the ability to detect damage has to be kept or increased when the real data set is applied.