• Proceedings of the Korean Information Science Society Conference
• /
• 2007.06c
• /
• pp.417-421
• /
• 2007

본 논문에서는 악보 영상을 인식하고 이를 연주할 수 있는 시스템을 제안한다. 이 시스템은 스캐너 또는 카메라로 악보 영상을 입력 받는다. 먼저 입력 영상을 전처리하여 영상 분석에 알맞은 형태로 변환시키고 선 추적으로 오선을 추출한 후, 템플릿 매칭을 이용해 음표, 쉼표, 보표, 조표 등을 추출하여 인식한다. 그리고 인식한 결과를 MIDI로 출력한다. 이 악보 인식 시스템을 통해 여러 실험 데이터를 검토해 본 바, 본 시스템이 실용적임을 보였다.

• Proceedings of the KIEE Conference
• /
• 2009.07a
• /
• pp.1835_1836
• /
• 2009

본 논문에서는 비접촉식 미세 측정 광학기에 의해 측정된 이미지를 Stitching 기술과 B-Spline 보간법에 의해 보다 빠르고, 정밀한 복원 기법을 제안한다. 이를 위해 먼저 각각의 이미지로부터 매칭, 피매칭 템플릿(Template)을 검출한다. 그런 다음 두 이미지의 오버랩(Overlap)되는 부분을 기준하여 기준면으로부터 roll, pitch, yaw 오차를 계산하여 매칭시킨다. 마지막으로, B-Spline 보간법에 의해 매칭된 부분을 연속화한다. 제안된 방법은 실험을 통해 그 응용 가능성을 증명한다.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2007.05a
• /
• pp.1074-1079
• /
• 2007

We will introduce speech/speaker recognition algorithm for the isolated word. In general case of speaker verification, Gaussian Mixture Model (GMM) is used to model the feature vectors of reference speech signals. On the other hand, Dynamic Time Warping (DTW) based template matching technique was proposed for the isolated word recognition in several years ago. We combine these two different concepts in a single method and then implement in a real time speaker/speech recognition system. Using our proposed method, it is guaranteed that a small number of reference speeches (5 or 6 times training) are enough to make reference model to satisfy 90% of recognition performance.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.11a
• /
• pp.295-298
• /
• 2009

웨이블렛 변환(Wavelet Transform)은 영상압축에 효율적인 방법으로 알려져 있으며 lifting scheme을 이용해서 쉽게 구현이 가능하다. 가장 널리 쓰이는 방법으로는 Daubechies 5/3 필터가 있고, 이를 바탕으로 하여 영상의 기하학적인 특성을 이용한 적응적 예측 방법이 많이 소개되었다. 본 논문에서는 적응적 예측을 위해 템플릿 매칭을 적용한 새로운 알고리즘을 제안하였다.

• Journal of Biosystems Engineering
• /
• v.23 no.4
• /
• pp.373-380
• /
• 1998

The contribution of this paper is to provide the methods for the production automation of potato microtuber using the vision process in growth monitoring. The first method deals with computation for the growth density in the primary growth process. The second method addresses cognition process to identify the number and the volume of potato microtuber in secondary growth process. The third is to decide whether potato microtubers are infected by a virus or bacteria in growth process. The computation for the growth density in the primary growth process uses the method of Labeling. The second and third methods use template matching based on color patterns. With the developed method using vision process, this experiment is capable of discriminating weekly growth-rate in primary growth process, 85% cognition rate in secondary process and identifying whether there are infections. Therefore, we conclude that our experimental results are capable of growth monitoring for mass production of potato microtubers.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.4
• /
• pp.420-425
• /
• 2009

In this paper, image detection and recognition algorithms are studied with respect to embedded carrier system. There are many suggested techniques to detect and recognize objects. But they have the propensity to need much calculation for high hit rate. Advanced and modified method needs to study for embedded systems that low power consumption and real time response are requested. The proposed methods were implemented using Intel(R) Open Source Computer Vision Library provided by Intel Corporation. And they run and tested on embedded system using a ARM920T processor by cross-compiling. They showed 1.6sec response time and 95% hit rate and supported the automated moving carrier system smoothly.

• Journal of Industrial Technology
• /
• v.28 no.B
• /
• pp.143-148
• /
• 2008

A real-time face detection is to find human faces robustly under the cluttered background free from the effect of occlusion by other objects or various lightening conditions. We propose a face detection system for real-time applications using wavelet decomposition method based on Gabor features. Firstly, skin candidate regions are extracted from the given image by skin color filtering and projection method. Then Gabor-feature based template matching is performed to choose face cadidate from the skin candidate regions. The chosen face candidate region is transformed into 2-level wavelet decomposition images, from which feature vectors are extracted for classification. Based on the extracted feature vectors, the face candidate region is finally classified into either face or nonface class by the Levenberg-Marguardt back-propagation neural network.

• 제어로봇시스템학회:학술대회논문집
• /
• 2002.10a
• /
• pp.100.3-100
• /
• 2002

This paper presents the method for detection and tracking of multiple humans robustly in mobile platform. The perception of human is performed in real time through the processing of images acquired from a moving stereo vision system. We performed multi-cue integration such as human shape, skin color and depth information to detect and track each human in moving background scene. Human shape is measured by edge-based template matching on distance transformed image. Improving robustness for human detection, we apply the human face skin color in HSV color space. And we could increase the accuracy and the robustness in both detection and tracking by applying random sampling stochastic estimati...

• Proceedings of the IEEK Conference
• /
• 1999.11a
• /
• pp.621-624
• /
• 1999

This paper describes a distance estimation method of object's motion in soccer image sequence by tracking field features. And we quantitatively evaluate the estimation accuracy We suppose that the input image sequence is taken with a camera on static axis and includes only zooming and panning transformation between frames. Adaptive template matching is adopted for non-rigid object tracking. For background compensation, feature templates selected from reference frame image are matched in following frames and the matched feature point pairs are used in computing Affine motion parameters. A perspective displacement field model is used for estimating the real distance between two position on Input Image. To quantitatively evaluate the accuracy of the estimation, we synthesized a 3 dimensional virtual stadium with graphic tools and experimented on the synthesized 2 dimensional image sequences. The experiment shows that the average of the error between the actual moving distance and the estimated distance is 1.84%.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2017.01a
• /
• pp.35-36
• /
• 2017

실시간 객체 인식은 카메라로부터 입력받은 영상 내에 존재하는 객체를 실시간으로 처리하는 기술로써 정확한 인식률과 빠른 인식 속도를 가져야 한다. 하지만 인식 속도가 보장되지 않으면 실시간으로 객체를 인식 할 수 없고 인식률이 보장되지 않으면 객체 인식을 통해 구현한 기능이 올바르게 동작하지 않을 수 도 있다. 따라서 본 논문에서는 실시간으로 객체를 인식하는 기술을 분류하고 연구 동향을 소개한다. 그리고 실시간 객체 인식을 위한 향후 연구 방향을 제시한다.