• Imaging Science in Dentistry
• /
• 제52권3호
• /
• pp.239-244
• /
• 2022

Purpose: Despite the proliferation of numerous morphometric and anthropometric methods for sex identification based on linear, angular, and regional measurements of various parts of the body, these methods are subject to error due to the observer's knowledge and expertise. This study aimed to explore the possibility of automated sex determination using convolutional neural networks(CNNs) based on lateral cephalometric radiographs. Materials and Methods: Lateral cephalometric radiographs of 1,476 Iranian subjects (794 women and 682 men) from 18 to 49 years of age were included. Lateral cephalometric radiographs were considered as a network input and output layer including 2 classes(male and female). Eighty percent of the data was used as a training set and the rest as a test set. Hyperparameter tuning of each network was done after preprocessing and data augmentation steps. The predictive performance of different architectures (DenseNet, ResNet, and VGG) was evaluated based on their accuracy in test sets. Results: The CNN based on the DenseNet121 architecture, with an overall accuracy of 90%, had the best predictive power in sex determination. The prediction accuracy of this model was almost equal for men and women. Furthermore, with all architectures, the use of transfer learning improved predictive performance. Conclusion: The results confirmed that a CNN could predict a person's sex with high accuracy. This prediction was independent of human bias because feature extraction was done automatically. However, for more accurate sex determination on a wider scale, further studies with larger sample sizes are desirable.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 추계학술대회
• /
• pp.229-232
• /
• 2022

최근 우리나라 연근해어업 생산이 줄어들고 있다. 2016년도 연근해어업 생산량이 44년 만에 100만톤 이하로 내려간 이후 회복이 되지 않고 줄어들고 있다. 이와 같은 수산자원 감소에 대응하기 위해 국제적으로 수산자원관리를 위하여 TAC(총허용어획량) 제도를 시행하고 있다. 우리나라는 1999년부터 TAC 제도를 도입하여 자원관리를 수행하고 있다. 본 논문에서는 TAC 제도 시행을 위해서 필수적인 육상 옵서버의 수산자원 조사에 활용이 가능한 인공지능 기반 어류 분류 및 무게 추정 시스템을 제안한다. 이 시스템은 라이다 센서가 탑재된 단말기를 이용하여 어류의 체장, 체고를 자동 측정 및 사진 촬영을 수행하는 앱과 클라우드 서버로 구성된다. 클라우드 서버에는 CNN 기반의 efficientnet 모델을 이용하여 어류 분류를 수행하고 자동 측정된 체장, 체고 정보를 이용하여 어류의 무게를 예측한다. 본 시스템을 이용하면 기존에 육상 옵서버가 위판장에서 줄자와 무게 측정 후 수기로 작성하는 기존 방식을 개선할 수 있다.

• Korean Journal of Radiology
• /
• 제23권3호
• /
• pp.343-354
• /
• 2022

Objective: To develop and evaluate a deep learning-based artificial intelligence (AI) model for detecting skull fractures on plain radiographs in children. Materials and Methods: This retrospective multi-center study consisted of a development dataset acquired from two hospitals (n = 149 and 264) and an external test set (n = 95) from a third hospital. Datasets included children with head trauma who underwent both skull radiography and cranial computed tomography (CT). The development dataset was split into training, tuning, and internal test sets in a ratio of 7:1:2. The reference standard for skull fracture was cranial CT. Two radiology residents, a pediatric radiologist, and two emergency physicians participated in a two-session observer study on an external test set with and without AI assistance. We obtained the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity along with their 95% confidence intervals (CIs). Results: The AI model showed an AUROC of 0.922 (95% CI, 0.842-0.969) in the internal test set and 0.870 (95% CI, 0.785-0.930) in the external test set. The model had a sensitivity of 81.1% (95% CI, 64.8%-92.0%) and specificity of 91.3% (95% CI, 79.2%-97.6%) for the internal test set and 78.9% (95% CI, 54.4%-93.9%) and 88.2% (95% CI, 78.7%-94.4%), respectively, for the external test set. With the model's assistance, significant AUROC improvement was observed in radiology residents (pooled results) and emergency physicians (pooled results) with the difference from reading without AI assistance of 0.094 (95% CI, 0.020-0.168; p = 0.012) and 0.069 (95% CI, 0.002-0.136; p = 0.043), respectively, but not in the pediatric radiologist with the difference of 0.008 (95% CI, -0.074-0.090; p = 0.850). Conclusion: A deep learning-based AI model improved the performance of inexperienced radiologists and emergency physicians in diagnosing pediatric skull fractures on plain radiographs.

• 전기학회논문지
• /
• 제61권11호
• /
• pp.1758-1764
• /
• 2012

A large number of wire rope has been used in various inderstiries as Cranes and Elevators from expanding the scale of the industrial market. But now, the management of wire rope is used as manually operated by rope replacement from over time or after the accident.It is caused to major accidents as well as economic losses and personal injury. Therefore its time to need periodic fault diagnosis of wire rope or supply of real-time monitoring system. Currently, there are several methods has been reported for fault diagnosis method of the wire rope, to find out the feature point from extracting method is becoming more common compared to time wave and model-based system. This method has implemented a deterministic modeling like the observer and neural network through considering the state of the system as a deterministic signal. However, the out-put of real system has probability characteristics, and if it is used as a current method on this system, the performance will be decreased at the real time. And if the random noise is occurred from unstable measure/experiment environment in wire rope system, diagnostic criterion becomes unclear and accuracy of diagnosis becomes blurred. Thus, more sophisticated techniques are required rather than deterministic fault diagnosis algorithm. In this paper, we developed the fault diagnosis of the wire rope using probability density estimation techniques algorithm. At first, The steady-state wire rope fault signal detection is defined as the probability model through probability distribution estimate. Wire rope defects signal is detected by a hall sensor in real-time, it is estimated by proposed probability estimation algorithm. we judge whether wire rope has defection or not using the error value from comparing two probability distribution.

• 제어로봇시스템학회논문지
• /
• 제22권11호
• /
• pp.912-918
• /
• 2016

This paper presents various strategies for humanoid vehicle driving and egress tasks. For driving, a tele-operating system that controls a robot based on a human operator's commands is built. In addition, an autonomous assistant module is developed for the operator. Normal position control can result in severe damage to robots when they egress from vehicles. To prevent this problem, another approach that mixes various joint control techniques is adopted in this study. Additionally, a footplate is newly designed and attached to the vehicle floor for the ground landing phase of the egress task. The attached plate enables the robot to step down onto the ground in a safe manner. For stable locomotion, a balance controller is designed for the humanoid. For the design of the controller, the robot is modeled using an inverted pendulum that consists of a spring and a damper. Then, a state feedback controller (with pole placement and a state observer) is built based on the simplified model. Many approaches that are presented in this paper were successfully applied to a full-sized humanoid, DRC-HUBO+, in the DARPA Robotics Challenge Finals, which were held in the United States in 2015.

• 전력전자학회논문지
• /
• 제12권1호
• /
• pp.65-73
• /
• 2007

본 논문은 1-Hz의 낮은 고정자 전류 주파수에서도 동작하는 순수적분 기반의 자속추정을 위한 온라인 전류측정 오차 보상방법을 제안한다. 오프셋 전류와 변환이득오차에 의한 역상분 전류 성분은 상태관측기를 이용하여 제거하고, 동시에 변환이득오차에 의한 역상분 전류는 동기좌표계에서 영구자석에 의하여 발생된 q축 자속을 기준모델에 의한 값과 추정된 자속에 의한 값 사이의 차이에 의하여 보상한다. 이 보상기는 PI제어기를 이용하여 두 값 사이의 오차가 0이 되도록 제어한다. 또한 적분기 초기값 오차 및 관측기의 전동기 상수 오차에 의한 잔여오차 보상방법도 제안하였다. 타당성을 입증하기 위하여 1.1-kW 영구자석형 동기전동기(PMSM)에 제안된 보상 방법을 구현하여 다양한 실험을 수행하였다.

• 천문학회보
• /
• 제43권2호
• /
• pp.44.4-45
• /
• 2018

(4179) Toutatis (Toutatis hereafter) is one of the Near-Earth Asteroids which has been studied most rigorously not only via ground-based photometric, spectroscopic, polarimetric, and radar observations, but also via the in-situ observation by the Chinese Chang'e-2 spacecraft. However, one of the most fundamental physical properties, the geometric albedo, is less determined. In order to derive the reliable geometric albedo and further study the physical condition on the surface, we made photometric observations of Toutatis near the opposition (i.e., the opposite direction from the Sun). We thus observed it for four days on 2018 April 7-13 using three 1.6-m telescopes, which consist of the Korean Microlensing Telescope Network (KMTNet). Since the asteroid has a long rotational period (5.38 and 7.40 days from Chang'e-2, Zhao et al., 2015), the continuous observations with KMTNet matches the purpose of our photometric study of the asteroid. The observed data cover the phase angle (Sun-asteroid-observer's angle) of 0.65-2.79 degree. As a result, we found that the observed data exhibited the magnitude changes with an amplitude of ~0.8 mag. We calculated the time-variable geometrical cross-section using the radar shape model (Hudson & Ostro 1995), and corrected the effect from the observed data to derive the geometric albedo. In this presentation, we will present our photometric results. In addition, we will discuss about the regolith particles size together with the polarimetric properties based on the laboratory measurements of albedo-polarization maximum.

• 한국콘텐츠학회논문지
• /
• 제10권8호
• /
• pp.186-197
• /
• 2010

최근 재현 장비의 목적은 장면을 정확하게 재현하는 것에서 사용자의 만족도를 높이는 것으로 변화하고 있다. 이는 카메라로 대표되는 재현 장비의 발전 방향이 성능 위주에서 품질 위주로, 개발자 중심에서 사용자 중심으로 전환되고 있다는 의미이다. 따라서 본 논문에서는 사용자의 인지 특성에 기반을 둔 사진의 화질 평가 방법을 제안하고자 한다. 사진은 다양한 물리적 화질 속성이 복합적으로 어우러져 완성되는 것으로 몇 가지 성능에 대한 평가만으로는 전반적인 화질을 평가할 수 없다. 따라서 제 3의 관찰자에 의한 주관적인 화질 평가가 필요하다. 주관적 화질 평가는 평가 방법과 결과 도출 방법이 쉬워야 하고, 도출된 평가 결과는 전체적인 화질을 어우르는 동시에 높은 신뢰도를 가져야 한다. 따라서 본 논문은 화질 만족도에 영향을 미치는 요소들은 분석하여 구체적인 평가 언어를 수립하고, 사진 이미지를 실험 자극으로 활용한 주관적 화질 평가 실험을 통해 화질 평가 항목을 수립함으로써 효율적이고 정확한 방법으로 주관적 화질을 평가할 수 있도록 하였다.

• 전자공학회논문지SC
• /
• 제43권4호
• /
• pp.35-44
• /
• 2006

LonWorks/IP 가상 디바이스 네트워크(VDN) 상의 전달지연은 실시간 분산제어 시스템의 성능과 안정성을 악화시킨다. LonWorks/IP VDN은 LonWorks 디바이스 네트워크와 IP( 데이터) 네트워크와의 통합네트워크이다. LonWorks/IP VDN 상에서의 서보제어를 수행할 경우 시간지연은 확률적인 특성을 강하게 나타낸다. 산업현장에 대한 예지보전을 위한 실시간 분산제어 환경에서 즉각적인 응답은 필수불가결한 요소이다. 그러므로 네트워킹 된 분산제어시스템의 안정성을 보장하고 성능을 개선하기 위해서는 시간에 따라 가변적인 불확실한 시간지연을 보상할 필요가 있다. 본 논문에서는 출력 되먹임 루프에 적절한 필터와 외란관측기를 이용한 제어기를 제안한다. 컴퓨터 모의실험을 통하여 제안된 제어기의 성능과 안정성이 Smith 예측기 기반의 내부모델제어기 (IMC)의 제어결과와 비교 제시된다. 제안된 제어기는 IMC 보다 안정성과 추종성능을 상당히 개선시킬 수 있으며 외란과 잡음에 강인한 특성을 갖는 것을 보인다. 그러므로 제안된 제어기는 가변적인 시간지연을 갖는 LonWorks/IP VDN 상에서 예지보전을 위한 실시간 분산제어에 매우 적합하다.

• Imaging Science in Dentistry
• /
• 제45권1호
• /
• pp.15-22
• /
• 2015

Purpose: This study was performed to assess the reproducibility of identifying the sella turcica landmark in a three-dimensional (3D) model by using a new sella-specific landmark reference system. Materials and Methods: Thirty-two cone-beam computed tomographic scans (3D Accuitomo$^{(R)}$ 170, J. Morita, Kyoto, Japan) were retrospectively collected. The 3D data were exported into the Digital Imaging and Communications in Medicine standard and then imported into the Maxilim$^{(R)}$ software (Medicim NV, Sint-Niklaas, Belgium) to create 3D surface models. Five observers identified four osseous landmarks in order to create the reference frame and then identified two sella landmarks. The x, y, and z coordinates of each landmark were exported. The observations were repeated after four weeks. Statistical analysis was performed using the multiple paired t-test with Bonferroni correction (intraobserver precision: p<0.005, interobserver precision: p<0.0011). Results: The intraobserver mean precision of all landmarks was <1 mm. Significant differences were found when comparing the intraobserver precision of each observer (p<0.005). For the sella landmarks, the intraobserver mean precision ranged from $0.43{\pm}0.34mm$ to $0.51{\pm}0.46mm$. The intraobserver reproducibility was generally good. The overall interobserver mean precision was <1 mm. Significant differences between each pair of observers for all anatomical landmarks were found (p<0.0011). The interobserver reproducibility of sella landmarks was good, with >50% precision in locating the landmark within 1 mm. Conclusion: A newly developed reference system offers high precision and reproducibility for sella turcica identification in a 3D model without being based on two-dimensional images derived from 3D data.