• 한국정밀공학회지
• /
• 제23권11호
• /
• pp.58-67
• /
• 2006

Due to the variety of signal processing and complicated mathematical analysis, it is not easy to accomplish 3D bin-picking with non-contact sensor. To solve this difficulties the reliable signal processing algorithm and a good sensing device has been recommended. In this research, 3D laser scanner and CCD camera is applied as a sensing device respectively. With these sensor we develop a two-step bin-picking method and reliable algorithm for the recognition of 3D bin object. In the proposed bin-picking, the problem is reduced to 2D intial recognition with CCD camera at first, and then 3D pose detection with a laser scanner. To get a good movement in the robot base frame, the hand eye calibration between robot's end effector and sensing device should be also carried out. In this paper, we examine auto-calibration technique in the sensor calibration step. A new thinning algorithm and constrained hough transform is also studied for the robustness in the real environment usage. From the experimental results, we could see the robust bin-picking operation under the non-aligned 3D hole object.

• 제어로봇시스템학회논문지
• /
• 제15권11호
• /
• pp.1122-1129
• /
• 2009

Recently taping robot with smart recognition function have been studied in the coil manufacturing field. Due to the difficulty of 3D surface processing from the complicated working environment, it is not easy to accomplish smart tape attachment motion with non-contact sensor. To solve these problems the applicable surface recognition algorithm and a flexible sensing device has been recommended. In this research, the fusion method between 1D displacement and 3D laser scanner is applied for robust tape attachment about cold rolled coil. With these sensors we develop a two-step exploration and the smart algorithm for the awareness of non-aligned coil's information. In the proposed robot system for tape attachment, the problem is reduced to coil's radius searching with laser displacement sensor at first, and then position and orientation detection with 3D laser scanner. To get the movement at the robot's base frame, the hand-eye compensation between robot's end effector and sensing device should be also carried out respectively. In this paper, we examine the auto-coordinate transformation method in the calibration step for the real environment usage. From the experimental results, it was shown that the taping motion of robot had a robust under the non-aligned cold rolled coil.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 1999년도 추계종합학술대회
• /
• pp.224-229
• /
• 1999

음성 인식 연구는 유사음소 단위의 인식시스템을 구축하는 방향과 단어 단위의 인식시스템에서의 효율을 최대화하는 방향으로 이루어지고 있다. 이중 유용한 유사음소 단위의 인식시스템 구현을 위해서는 음소의 경계 검출 문제와 검출된 음소에 대한 인식률 향상 문제가 해결되어야 한다. 기존의 LPC(Linear Predictive Coefficient) 방법들은 기준 음소데이터의 LPC와 입력 음성프레임의 LPC 사이의 거리를 Itakura-Saito 방법으로 구하여 음소의 경계를 검출하였으며, 근래에는 MFCC(Mel-Frequency-Cepstrum Coefficient)를 이용하여 스펙트럼의 천이부분을 음소의 경계로 검출하는 방법들이 제안되어왔으나 이러한 방법들은 공통적으로 적응성이 미비하다는 단점이 있다. 본 논문에서는 이러한 단점을 극복하기 위해 음소경계검출을 위해서는 auto-correlation을 이용하고 음소인식을 위해서는 적응성이 뛰어난 다층 Feed-Forward 신경망을 사용하는 새로운 인식시스템을 제안하였다 제안하는 시스템은 기존의 방법들보다 적응성이 뛰어나고 특징추출부분과 인식 부분의 알고리듬이 독립적이라는 장점을 가지며 프레임단위의 음소인식시스템의 구현 가능성을 확인해 주었다.

• 핵의학기술
• /
• 제21권2호
• /
• pp.55-64
• /
• 2017

최근 SPECT/CT의 보급과 함께 다양한 영상보정 방법들을 빠르고 정확하게 적용할 수 있게 되면서, 영상품질 향상과 더불어 정량적 정확성까지 기대할 수 있게 되었다. 그중 Collimator Detector Response (CDR) 회복(recovery)은 검출기면의 거리로부터 발생된 blurring 효과를 보상하여 분해능 회복을 목적으로 하는 보정방법이다. 본 연구에서는 SPECT/CT 영상에서 CDR recovery 가 적용되었을 때 검출거리 변화에 따른 정량적 변화를 알아보고자 하였다. 검출거리의 변화에 따른 획득 계수의 차이를 알아보고자 검출거리를 궤도방식(obit type)에 따라 Circular는 X, Y축 반경 30 cm, Non-Circular는 X, Y축 반경 21 cm, 10 cm, Non-Circular Auto(=Auto Body Contouring_ ABC, spacing limit 1 cm)로 설정하였고, 재구성 방법은 CDR recovery(CDRr)의 사용 유/무에 따른 계수 회복 차이를 알아보고자 OSEM (w/o CDRr)와 Astonish(3D-OSEM with CDRr)로 구분하여 적용하였다. 이 때 감쇠, 산란, 붕괴 보정은 모든 영상에 공통 적용하였다. 정량적 평가를 위해 교정인자(calibration factor_CF) 산출을 목적으로 교정영상(cylindrical phantom, $^{99m}TcO_4$ 123.3 MBq, 물 9293 ml)을 획득하였고, 팬텀 실험을 위하여 50 cc 주사기에 물 31 ml를 채우고 $^{99m}TcO_4$ 123.3 MBq를 설정하여 팬텀영상을 획득하였다. 팬텀 영상에서 주사기 전체 체적에 VOI(volume of interest)를 설정하여 각 조건별로 총 계수 값을 측정하였고, CF를 적용시켜 설정된 참값 대비 추정값의 오차를 구하여 보정에 따른 정량적 정확성을 확인하였다. 산출된 CF는 154.27 (Bq/ml/cps/ml)이며, 각 조건별 영상에서 참값 대비 추정값은 OSEM에서 Circular 86.5%, Non-Circular 90.1%, ABC 91.3% Astonish에서 Circular 93.6%, Non-Circular 93.6%, ABC 93.9%으로 분석되었다. OSEM은 검출거리가 가까울수록 정확성이 높아졌으며, Astonish의 경우에는 거리와 상관없이 거의 유사한 값을 나타내었다. 오차는 OSEM Circular(-13.5%)에서 가장 크고, Astonish ABC(-6.1%)에서 가장 적었다. SPECT/CT영상에서 CDR recovery 적용을 통한 거리보상이 이루어 졌을 때 검출거리가 먼 조건에서도 근접검출과 거의 동일한 정량적 정확성을 보였고, 검출거리의 변화에 영향을 받지 않고 정확한 보정이 가능한 것을 확인 할 수 있었다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2006년도 하계종합학술대회
• /
• pp.825-826
• /
• 2006

In this paper, we propose an algorithm that applies Rolle's theorem to automatically detect and label peak III and V of the normal, suprathreshold auditory brainstem response (ABR). ABR waveform were recorded from 55 normal-hearing ears at screening levels varying from 30 to 60 dBnHL. For each ABR waveform, the peak-finding algorithm proceeded in fourth steps: (1) Select maximum and minimum values of the target ABR waveform, (2) divide this range into n equal parts, (3) effective candidate peaks in the ABR waveform are identified using Rolle's theorem (4) peak III and V are identified from these candidate peaks based on their latency and morphology. As a result, proposed auto dectection method showed high correlation and accuracy with manual detection method performed by clinician. By using proposed algorithm, clinician can detect and label peak III and V faster and more efficient than manual detection method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권9호
• /
• pp.2927-2941
• /
• 2022

The primary task of machine fault diagnosis is to judge whether the current state is normal or damaged, so it is a typical binary classification problem with mutual exclusion. Mutually exclusive events and out-of-domain detection have one thing in common: there are two types of data and no intersection. We proposed a fusion model method to improve the accuracy of machine fault diagnosis, which is based on the mutual exclusivity of events and the commonality of out-of-distribution detection, and finally generalized to all binary classification problems. It is reported that the performance of a convolutional neural network (CNN) will decrease as the recognition type increases, so the variational auto-encoder (VAE) is used as the primary model. Two VAE models are used to train the machine's normal and fault sound data. Two reconstruction probabilities will be obtained during the test. The smaller value is transformed into a correction value of another value according to the mutually exclusive characteristics. Finally, the classification result is obtained according to the fusion algorithm. Filtering normal data features from fault data features is proposed, which shields the interference and makes the fault features more prominent. We confirm that good performance improvements have been achieved in the machine fault detection data set, and the results are better than most mainstream models.

• 한국통신학회논문지
• /
• 제32권3C호
• /
• pp.189-194
• /
• 2007

본 논문에서는 시간 동기 추정을 위한 잡음 예측이 불완전한 경우에 직교분할 주파수 다중화 (OFDM) 시스템의 시간 동기 확률을 다중 경로 레일레이 페이딩 채널 환경에서 분석한다. 시간 동기 방식은 반복되는 CAZAC 시퀀스에 대한 상관기 구조를 가지며, 심벌간 간섭을 줄이기 위한 pre-advancement 기법을 적용한다. 시스템의 성능 평가 척도로서, 다중 경로 레일레이 채널 환경에서 유도된 동기 검파 확률을 분석한다.

• 대한의용생체공학회:의공학회지
• /
• 제17권2호
• /
• pp.241-246
• /
• 1996

This paper describes a new system that detects radial pulse wave and allows the diagnosis of malfunctions of cardiovascular system by analyzing the waveforms with the newly proposed algorithm. The system consists of a sensor part and a data processing part within which a new detection algorithm is incorporated In acquiring radial pulse signal noninvasively, the sensor used in this system is a new combinational fiber-optic sensor which has a detecting Part and a transmitting Part. Also, In order to analyze the characteristics of pulsation quantitatively, the algorithm proposed in this paper is a method that runs in parallel with both the data of ECG and differential pulse simultaneously. these concepts are based upon the idea that thfee Q points of ECG give obious discrimination of one entire period of pulse in any abnormal cases, and newly defined feature lines at the differential counterpart can be used to recogrlize sDme significant points in one period of pulses.

• Journal of Information Processing Systems
• /
• 제15권1호
• /
• pp.100-115
• /
• 2019

Malicious code distribution on the Internet is one of the most critical Internet-based threats and distribution technology has evolved to bypass detection systems. As a new defense against the detection bypass technology of malicious attackers, this study proposes the automated tracing of malicious websites in a malware distribution network (MDN). The proposed technology extracts automated links and classifies websites into malicious and normal websites based on link structure. Even if attackers use a new distribution technology, website classification is possible as long as the connections are established through automated links. The use of a real web-browser and proxy server enables an adequate response to attackers' perception of analysis environments and evasion technology and prevents analysis environments from being infected by malicious code. The validity and accuracy of the proposed method for classification are verified using 20,000 links, 10,000 each from normal and malicious websites.

• Structural Monitoring and Maintenance
• /
• 제3권4호
• /
• pp.395-406
• /
• 2016

Modal parameter identification has received much attention recently for their usefulness in earthquake engineering, damage detection and structural health monitoring. The identification method based on Matrix Pencil technique is adopted in this paper to identify structural modal parameters, such as natural frequencies, damping ratios and modal shapes using impulse vibration responses. This method can also be applied to dynamic responses induced by stationary and white-noise inputs since the auto- and cross-correlation function of the two outputs has the same form as the impulse response dynamic functions. Matrix Pencil method is very robust to noise contained in the measurement data. It has a lower variance of estimates of the parameters of interest than the Polynomial Method, and is also computationally more efficient. The numerical simulation results show that this technique can identify modal parameters accurately even if the noise level is high.