• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.287-288
• /
• 2006

CG and Photo-realistic image composition in the ocean scenes is frequently used in movies and TV advertisement. But it is very difficult task because it's impossible to use calibration tool in outdoor environment or to use auto-calibration algorithm using natural features like KLT(Kanade Lucas Tomasi feature tracker) from the ocean scene. We propose a simple, effective method for solving camera motion using previous knowledge about background structure. We applied our method to the production of a commercial movie, 'Hanbando' and the result was satisfactory.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.5
• /
• pp.473-479
• /
• 2009

This paper proposes an efficient method to locate the automated guided vehicle (AGV) into a specific parking position using artificial visual landmark and vision-based algorithm. The landmark has comer features and a HSI color arrangement for robustness against illuminant variation. The landmark is attached to left of a parking spot under a crane. For parking, an AGV detects the landmark with CCD camera fixed to the AGV using Harris comer detector and matching descriptors of the comer features. After detecting the landmark, the AGV tracks the landmark using pyramidal Lucas-Kanade feature tracker and a refinement process. Then, the AGV decreases its speed and aligns its longitudinal position with the center of the landmark. The experiments showed the AGV parked accurately at the parking spot with small standard deviation of error under bright illumination and dark illumination.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.372-374
• /
• 2003

The optical flow theory can be utilized for automatically tracking and positioning homologous points in digital video (DV) image sequences. In this paper, the Lucas-Kanade optical flow estimation (LKOFE) method and the normalized cross-correlation (NCC) method are compared and analyzed using the DV image sequences acquired by our SONY DCRPC115 DV camera. Thus, an improved optical flow estimation procedure, called 'Iterated Optical Flow Estimation (IOFE)', is presented. Our test results show that the trackable range of 3${\sim}$4 pixels in the LKOFE procedure can be apparently enlarged to 30 pixels in the IOFE.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2014.04a
• /
• pp.781-784
• /
• 2014

Face recognition under controlled settings, such as limited viewpoint and illumination change, can achieve good performance nowadays. However, real world application for face recognition is still challenging. In this paper, we use Affine SIFT to detect affine invariant local descriptors for face recognition under large viewpoint change. Affine SIFT is an extension of SIFT algorithm. SIFT algorithm is scale and rotation invariant, which is powerful for small viewpoint changes in face recognition, but it fails when large viewpoint change exists. In our scheme, Affine SIFT is used for both gallery face and probe face, which generates a series of different viewpoints using affine transformation. Therefore, Affine SIFT allows viewpoint difference between gallery face and probe face. Experiment results show our framework achieves better recognition accuracy than SIFT algorithm on FERET database.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2012.11a
• /
• pp.594-596
• /
• 2012

본 논문에서는 손 모션에 대하여 피부색 검출을 기반으로 전역적인 모션을 추적하고 모션 벡터를 생성하여 제스처를 인식하는 방법을 제안한다. 추적을 위하여 Shi-Tomasi 특징점 검출 방법과 Lucas-Kanade 옵티컬 플로우 추정 방법을 사용한다. 손 모션을 추적하는 경우 손의 모양이 다양하게 변화하므로 초기에 검출된 특징점을 계속적으로 추적하는 일반적인 방법으로는 손의 모션을 제대로 추적할 수 없다. 이에 본 논문에서는 프레임마다 새로운 특징점을 검출한 후 옵티컬 플로우를 추정하고 이상치(outlier)를 제거하여 손 모양의 변화에도 추적을 통한 모션 벡터 생성이 가능하도록 한다. 모션 벡터들로 인공 신경망을 사용한 판별 과정을 수행하여 최종적으로 손 모션 제스처에 대한 인식이 가능하도록 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.11a
• /
• pp.108-111
• /
• 2008

본 논문에서는 KLT(Kanade-Lucas-Tomasi) 추적기에서 특징점의 개수를 일정하게 유지시키기 위해 존재하는 특징점의 관리 부분을 FPGA(Field Programmable Gate Array)에 구현하기 위한 구조를 제안한다. FPGA 에 구현하기 위해 한정된 자원을 효과적으로 사용하도록 하는 것을 목표로 연산량이 많은 부분을 적은 연산량으로 구현 가능한 것으로 대체하고, 메모리의 크기와 접근 회수를 줄이기 위한 방법을 고려했다. 구현이 간단한 Harris 코너 검출기를 이용하여 특징점을 선택하고, 나눗셈 연산이 필요 없는 히스토그램을 이용하여 임계값을 설정해 특징점을 관리했다. C 언어로 시뮬레이션을 수행하여 제안한 방법을 확인했고, 기존의 특징점 관리 방법과의 비교를 통해 검증했다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.11a
• /
• pp.90-92
• /
• 2008

컴퓨터 비전 시스템에서 장면간의 객체 추적은 매우 유용한 도구이다. 이러한 추적 문제를 손쉽게 해결하기 위하여, 많이 알려진 Kanade-Lucas-Tomasi(KLT)는 6 개의 인자를 사용한 시스템을 설계를 하였다. 그러나, 현실적으로 많은 지역 영역(local patch)을 추적하는데 있어서는 4 개의 인자 (수평과 수직방향 이동, 균등 비례적 축소, 회전)으로 충분히 설명이 가능할 수 있다. 본 실험에서는 이 4 개의 인자로 정의되는 시스템을 새롭게 정의하고, 실제적인 KLT 와 비교실험을 하였다. 실험결과 적은 수의 인자로 설명하였음에도 불구하고, KLT 보다 좋은 성능을 나타냈다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39C no.10
• /
• pp.948-956
• /
• 2014

Today, Intelligent Vehicle Detection System seeks to reduce the negative factors, such as accidents over to get the traffic information of existing system. This paper proposes detection algorithm for the illegal U-turn vehicles which can cause critical accident among violations of road traffic laws. We predicted that if calculated optical flow vectors were shown on the illegal U-turn path, they would be cause of the illegal U-turn vehicles. To reduce the high computational complexity, we use the algorithm of pyramid Lucas-Kanade. This algorithm only track the key-points likely corners. Because of the high computational complexity, we detect center lane first through the color information and progressive probabilistic hough transform and apply to the around of center lane. And then we select vectors on illegal U-turn path and calculate reliability to check whether vectors is cause of the illegal U-turn vehicles or not. Finally, In order to evaluate the algorithm, we calculate process time of the type of algorithm and prove that proposed algorithm is efficiently.

• The KIPS Transactions:PartB
• /
• v.14B no.7
• /
• pp.493-502
• /
• 2007

An on-line registration technique is presented to register multi-view range images for the 3D reconstruction of real objects. Using a range camera, we first acquire range images and photometric images continuously. In the range images, we divide object and background regions using a predefined threshold value. For the coarse registration of the range images, the centroid of the images are used. After refining the registration of range images using a projection-based technique, we use a modified KLT(Kanade-Lucas-Tomasi) tracker to match photometric features in the object images. Using the modified KLT tracker, we can track image features fast and accurately. If a range image fails to register, we acquire new range images and try to register them continuously until the registration process resumes. After enough range images are registered, they are integrated into a 3D model in offline step. Experimental results and error analysis show that the proposed method can be used to reconstruct 3D model very fast and accurately.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.6B
• /
• pp.930-939
• /
• 2010

Recently, many researches for making more comfortable input device based on gaze detection have been vigorously performed in human computer interaction. However, these previous researches are difficult to be used in IPTV environment because these methods need additional wearing devices or do not work at a distance. To overcome these problems, we propose a new way of controlling IPTV interface by using a detected face and eye positions in single static camera. And although face or eyes are not detected successfully by using Adaboost algorithm, we can control IPTV interface by using motion vectors calculated by pyramidal KLT (Kanade-Lucas-Tomasi) feature tracker. These are two novelties of our research compared to previous works. This research has following advantages. Different from previous research, the proposed method can be used at a distance about 2m. Since the proposed method does not require a user to wear additional equipments, there is no limitation of face movement and it has high convenience. Experimental results showed that the proposed method could be operated at real-time speed of 15 frames per second. Wd confirmed that the previous input device could be sufficiently replaced by the proposed method.