• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.5
• /
• pp.539-544
• /
• 2004

We investigate a visual object tracking algorithm based upon particle filters, namely CONDENSATION, in order to combine multiple observation models such as active contours of digitally subtracted image and the particle measurement of object color. The former is applied to matching the contour of the moving target and the latter is used to independently enhance the likelihood of tracking a particular color of the object. Particle filters are more efficient than any other tracking algorithms because the tracking mechanism follows Bayesian inference rule of conditional probability propagation. In the experimental results, it is demonstrated that the suggested contour tracking particle filters prove to be robust in the cluttered environment of robot vision.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.18 no.5
• /
• pp.171-177
• /
• 2018

Recently there are many image datasets which has variety of data class and point to extract general features. But in order to this variety data class and point, deep learning model trained this dataset has not good performance in heterogeneous data feature local area. In this paper, we propose the structure which use sub-category and openset object detection methods to train more robust model, named multi-branch tree using ASSL. By using this structure, we can have more robust object detection deep learning model in heterogeneous data feature environment.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.5
• /
• pp.203-216
• /
• 2004

Recent advances in video coding technology have resulted in rapid growth of application in mobile communication, With this explosive growth reliable transmission and error resilient technique become increasingly necessary to offer high quality multimedia service. In this paper, we present the result of our investigation on the error resilient performance evaluation of the MPEG-4 simple profile under the H.324/M and the H.264/AVC baseline under the IP packet networks. Especially, we have tested error resilient tools of MPEG-4 simple profile such as resynchronization marker insertion, data partitioning, and of H.264/AVC baseline such as the flexible macroblock ordering (FMO) scheme. The objective quality of decoded video is measured in terms of rate and PSNR under various random bit and burst error conditions.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.39 no.4
• /
• pp.389-394
• /
• 2002

In this paper, we propose a robust stereo matching algorithm using nonlinear extended Kalman filter. The proposed algorithm estimates disparity using nonlinear extended Kalman filter and compares left image to right image for obtained disparity. As this process is run iteratively, we get disparity only with a few search. And, we can get robust stereo matching results by comparing left image to right image using bilinear interpolation to consider influence of neighborhood pixel. We compared SSD algorithm which is widely used, in stereo matching method, to result of the proposed algorithm. As the result, the proposed algorithm has an outstanding matching performance.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.8
• /
• pp.590-596
• /
• 2004

This paper Presents a new self-adaptation algorithm based on maximum a posteriori (MAP) eigenvoice for Korean connected digit recognition. The proposed MAP eigenvoice is developed by introducing a probability density model for the eigenvoice coefficients. The Proposed approach provides a unified framework that incorporates the Prior model into the conventional eigenvoice estimation. In self-adaptation system we use only one adaptation utterance that will be recognized, we use MAP eigenvoice that is most robust adaptation. In series of self-adaptation experiments on the Korean connected digit recognition task. we demonstrate that the performance of the proposed approach is better than that of the conventional eigenvoice algorithm for a small amount of adaptation data.

• 제어로봇시스템학회:학술대회논문집
• /
• 1986.10a
• /
• pp.503-505
• /
• 1986

주어진 제어대상 모델에 대하여 제어기를 구성하여 실제로 적용하는 경우 모델의 불일치, 모델링에서 고려하지 않은 외란(disturbance), 측정잡음등에 의하여 성능이 설계시와 달라진다. 실제적용에서도 성능을 계속 유지하기 위하여 제어기는 안정성, 계수변화(parameter variation)에 대한 강인성(robustness), 외란상쇄(disturbance rejection) 및 측정잡음에 둔감함등의 특성을 가져야 한다. 귀환(feedback)을 사용하여 제어기를 구성하는 경우 위의 모든 조건을 만족 시킬 수 없으므로 제어목적에 따라 적당한 조건을 선정하여 중요한 특성을 주로 갖게 한다. 본 논문에서는 쌍동선(small waterplane area twin hull ship-SWATHS)에 대하여 PID, LQ, LQG 제어기를 구성하여 안정성, 계수 변화에 대한 강인성, 외란 상쇄 및 측정잡음의 영향을 비교하였다. 쌍동선의 경우 다른 단동선(mono hull ship)에 비하여 접수면(waterplane)이 적으므로 무게증변을 흡수할 수 있는 복원력이 약하여 적은 외력에도 상하동요(heave)와 종동요(pitch)가 심하게 일어난다. 이러한 동요를 줄이는 것이 쌍동선의 제어목적이다. 본 연구에서는 먼저 선형화된 수직축 운동방정식을 이용하여 선체운동의 모델을 구했으며 중첩의 원리(super-position theorem)에 의하여 주파수 응답의 합으로 파도입력을 모델링 하였으며 제어를 위하여 필요한 측정치는 IMU(Inertial Measurement Unit)에서 제공된다고 가정하였다. 쌍동선의 동요의 원인은 파도, 바람, 조류 등이 있으나 파도에 의한 영향이 가장 크므로 본 논문에서는 파도에 의한 영향만을 고려하였다. 파도는 쌍동선에 외란으로 작용하며 측정할 수 없는 양이므로 PID, LQ 제어에서는 제어모델에 포함되지 않지만 LQG 제어에서는 제어모델에 포함된다. LQG 제어의 경우 제어모델에 파도를 백색잡음으로 가정하고 제어기를 구성한 것 (LQG1)과 2차의 쉐이핑필터(shaping filter)를 사용하여 구성한 것(LQG2)으로 나누었다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.39 no.1
• /
• pp.40-48
• /
• 2002

Three-Dimensional scene reconstruction from images acquired with different viewpoints is possible as estimating Fundamental matrix(F-matrix) that indicates the epipolar geometry of two images. Correspondence points required to calculate F-matrix of two images include noise such as miss matches, so generally it is hard to calculate F-matrix accurately. In this paper, we classify noise into two types; outlier and minute noise. we propose SSOR algorithm that estimate F-matrix effectively. SSOR algorithm is rejecting outlier step by step in a noise environment. To evaluate the performance of proposed algorithm we simulated with synthetic images and real images. As a result of simulation we show that proposed algorithm is better than conventional algorithms.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.1
• /
• pp.169-174
• /
• 2010

This paper explains end-point detection method for better speech recognition rates. The proposed method determines speech and non-speech region with the entropy and the harmonic detection of speech. The end-point detection using entropy on the speech spectral energy has good performance at the high SNR(SNR 15dB) environments. At the low SNR environment(SNR 0dB), however, the threshold level of speech and noise varies, so the precise end-point detection is difficult. Therefore, this paper introduces the end-point detection methods which uses speech spectral entropy and harmonics. Experiment shows better performance than the conventional entropy methods.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.163-166
• /
• 2001

초음파를 이용하여 인체 조직의 특성을 알아내는 방법은 매우 광범위하게 응용되어 오고있다. 그 중에서 초음파를 발생시킨 후 반사되어 되돌아오는 신호를 측정하여 그 감쇠 정도로부터 조직의 특성을 추정하는 방법이 많이 사용되고 있다. 이러한 감쇠현상은 발생된 초음파가 조직 내에서 흡수 또는 산란현상을 거치면서 주파수가 변이를 일으키기 때문에 발생한다. 따라서, 조직의 감쇠 특성을 알아내기 위해서, 주파수의 함수로 근사할 수 있는 감쇠 계수(attenuation coefficient)를 이용하여 시간에 따라 달라지는 주파수 변화를 추정한다. 그러나, 기존의 Ah(Auto-Regressive) 모델을 통한 시간영역 및 주파수 영역에서의 추정 방법을 사용하면 잡음이 존재하는 상황에서 시변 신호를 추정하는데 성능이 많이 저하된다. 본 논문에서는 이러한 단점을 보완하기 위해서, 가변 망각 인자와 재귀적인 TLS(Total Least Squares) 방법을 사용하여 시간에 따라 변하는 신호를 정확하게 추정하고 잡음환경에도 강인한 알고리듬을 제안하였다. 또한, 제안된 알고리듬은 추정 성능을 향상시킬 뿐 아니라 감쇠정도의 강약에 관계없이 망각인자의 값을 적응적으로 변화시켜 동작하는 장점을 가진다.

• The Journal of the Acoustical Society of Korea
• /
• v.19 no.3
• /
• pp.45-50
• /
• 2000

This paper proposes a delta energy method based on Parameter Filtering(PF), which is a speech segmentation algorithm for text dependent speaker verification system over telephone line. Our parametric filter bank adopts a variable bandwidth along with a fixed center frequency. Comparing with other methods, the proposed method turns out very robust to channel noise and background noise. Using this method, we segment an utterance into consecutive subword units, and make models using each subword nit. In terms of EER, the speaker verification system based on whole word model represents 6.1%, whereas the speaker verification system based on subword model represents 4.0%, improving about 2% in EER.