• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.5
• /
• pp.1217-1230
• /
• 2010

The dynamic programming based methods, a kind of globally optimizing stereo matching methods, has the inherent advantage that the occlusion regions can be found during the process. But it also has a serious drawback of streaking phenomena. This paper focuses on reducing the streaking phenomena by adjusting the penalties in calculating the cost matrix and re-establishing the optimal path in the back-tracing process using the boundary information of the image. Especially we use a pixel expansion method in re-establishing the path, which is the results from expanding the pixel information of the ones just left the boundaries. Experiments with the four image pairs provided by the Middlebury site showed the results that the proposed method has the disparity error ratio of 6.33% and the rank is 29, which is competitive to the best method among the previously published dynamic programming based methods.

• Journal of Korean Society for Geospatial Information Science
• /
• v.7 no.2 s.14
• /
• pp.47-56
• /
• 1999

In this study, tie point observation in aerial triangulation was automated by the image processing methods. The technique includes boundary extraction and We matching processes. The procedures were applied to extract points of Interest and to find their conjugate points in the other images. The image coordinates of the identified points were then used to compute their absolute coordinates. An algorithm was developed in this study for the automation of observation in aerial triangulation, which is a manual process of selecting a tie point and recording the image coordinate of the selected point. The developed algorithm automates this process through the application of a mathematical operator to extract points of interest from an arbitrary image. The root m square error of image coordinates of the developed algorithm is $6.8{\mu}m$, which is close to that of the present analytical method. In a manual environment, the accuracy of the result of a photogrammetric process is heavily dependant on the level of skill and experience of the human operator. No such problem exists in an automated system. Also, as a result of the automated system, the time spent in the observation process could be reduced by a factor of 61.2%, thereby reducing the overall cost.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.3
• /
• pp.407-415
• /
• 2012

In this paper, we propose a new stereo matching algorithm by using color information especially for stereo images containing human beings in the applications such as tele-presence system. In the proposed algorithm, we first remove the background regions by using a threshold value for stereo images obtained by stereo camera and then find an initial disparity map and segment a given image into R, G, B and white color components. We also obtain edges in the segmented image and estimate the disparity from the extract boundary regions. Finally, we generate the final disparity map by properly combining the disparity map of each color component. Experiment results show better performance compared with the window based method and the dynamic programing method especially for stereo images with human being.

• Proceedings of the KSRS Conference
• /
• 2006.03a
• /
• pp.169-172
• /
• 2006

본 논문에서는 1m 해상도의 위성영상으로부터 건물의 경계선을 검출하기 위해 영상분할과 변이지도(disparity map)를 이용하는 새로운 방법을 제안한다. Watershed 방법으로 영상을 분할하고 분할된 영역 내부의 변이를 다중정합창틀(multiple matching window)과 결합된 다차원특징벡터정합(multi-dimensional feature vector matching)을 이용하여 계산한다 분할된 인접 영역들 가운데 panchromatic 및 multispectral 밝기값과 변이의 평균값이 유사하면 두 영역을 결합하여 하나의 영역을 생성하고 이 과정을 반복적으로 수행한다. 영역의 평균 변이값이 기준 값보다 크면 이를 건물 지붕 영역으로 결정한다. IKONOS 위성영상에 제안한 방법을 적용하여 작은 건물이 밀집되어 있는 도시 지역에서 건물 지붕의 영역과 경계선을 효과적으로 검출할 수 있었다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.01a
• /
• pp.71-75
• /
• 2009

In this paper, we present a real-time method to detect moving objects in a rotating and zooming camera. It is useful for camera surveillance of fixed but rotating camera, camera on moving car, and so on. We first compensate the global motion, and then exploit the displaced frame difference (DFD) to find the block-wise boundary. For robust detection, we propose a kind of image to combine the detections from consecutive frames. We use the block-wise detection to achieve the real-time speed, except the pixel-wise DFD. In addition, a fast block-matching algorithm is proposed to obtain local motions and then global affine motion. In the experimental results, we demonstrate that our proposed algorithm can handle the real-time detection of common object, small object, multiple objects, the objects in low-contrast environment, and the object in zooming camera.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.805-807
• /
• 2005

본 논문에서는 동적 계획법(dynamic programming)과 이미지 세그먼트(segment)를 이용한 새로운 스테레오 정합(stereo matching)기법을 제안한다. 일반적으로 동적 계획법(dynamic programming)은 빠르면서도 비교적 정확하고, 조밀(dense)한 disparity map을 얻을 수 있다. 그러나 경계(boundary)근처의 폐색지역(occlusion region)이나, 텍스쳐가 적은 모호한 영역에서는 잘못된 결과를 유도할 수 있다. 본 논문에서는 이러한 문제점들을 해결하기 위해 먼저 이미지를 아주 작은 영역으로 분할(over-segmentation)하고, 이런 작은 영역들이 비슷한 disparity를 가질 것이라고 가정한다. 다음으로 동적 계획법(dynamic programming)을 통해 정합을 수행한다. 여기서 계산비용(cost)은 기존의 정합윈도우 안에서 세그먼트 영역을 적용한 새로운 비용함수를 사용하며, 이 새로운 비용함수를 통해 정확도를 높인다. 마지막으로 동적 계획법을 통하여 얻어진 조밀한 disparity map을 세그먼트 영역들의 시각특성(visibility)과 유사도(similarity)를 이용하여 에러를 찾아내고, 세그먼트 정합을 통해 수정함으로 정확한 disparity map을 찾아낸다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.23 no.11 s.170
• /
• pp.1929-1939
• /
• 1999

This paper presents a neural network approach, which was named PRONET, to 3D object recognition and pose calculation. 3D objects are represented using a set of centroidal profile patterns that describe the boundary of the 2D views taken from evenly distributed view points. PRONET consists of the training stage and the execution stage. In the training stage, a three-layer feed-forward neural network is trained with the centroidal profile patterns using an error back-propagation method. In the execution stage, by matching a centroidal profile pattern of the given image with the best fitting centroidal profile pattern using the neural network, the identity and approximate orientation of the real object, such as a workpiece in arbitrary pose, are obtained. In the matching procedure, line-to-line correspondence between image features and 3D CAD features are also obtained. An iterative model posing method then calculates the more exact pose of the object based on initial orientation and correspondence.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.4
• /
• pp.113-120
• /
• 2004

In this paper, we propose an efficient feature extraction and image retrieval algorithm for content-based retrieval method. First, we extract the object using Gaussian edge detector for input image which is key frames of MPEG video and extract the object features that are location feature, distributed dimension feature and invariant moments feature. Next, we extract the characteristic color feature using the proposed HAQ(Histogram Analysis md Quantization) algorithm. Finally, we implement an retrieval of four features in sequence with the proposed matching method for query image which is a shot frame except the key frames of MPEG video. The purpose of this paper is to propose the novel content-based image retrieval algerian which retrieves the key frame in the shot boundary of MPEG video belonging to the scene requested by user. The experimental results show an efficient retrieval for 836 sample images in 10 music videos using the proposed algorithm.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.58 no.5
• /
• pp.1025-1034
• /
• 2009

In steel making production line, steel slabs are given a unique identification number. This identification number, Slab management number(SMN), gives information about the use of the slab. Identification of SMN has been done by humans for several years, but this is expensive and not accurate and it has been a heavy burden on the workers. Consequently, to improve efficiency, automatic recognition system is desirable. Generally, a recognition system consists of text localization, text extraction, character segmentation, and character recognition. For exact SMN identification, all the stage of the recognition system must be successful. In particular, the text localization is great important stage and difficult to process. However, because of many text-like patterns in a complex background and high fuzziness between the slab and background, directly extracting text region is difficult to process. If the slab region including SMN can be detected precisely, text localization algorithm will be able to be developed on the more simple method and the processing time of the overall recognition system will be reduced. This paper describes about the slab region localization using SIFT(Scale Invariant Feature Transform) features in the image. First, SIFT algorithm is applied the captured background and slab image, then features of two images are matched by Nearest Neighbor(NN) algorithm. However, correct matching rate can be low when two images are matched. Thus, to remove incorrect match between the features of two images, geometric locations of the matched two feature points are used. Finally, search rectangle method is performed in correct matching features, and then the top boundary and side boundaries of the slab region are determined. For this processes, we can reduce search region for extraction of SMN from the slab image. Most cases, to extract text region, search region is heuristically fixed [1][2]. However, the proposed algorithm is more analytic than other algorithms, because the search region is not fixed and the slab region is searched in the whole image. Experimental results show that the proposed algorithm has a good performance.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.01a
• /
• pp.273-277
• /
• 2009

A compressed video stream is very sensitive to transmission errors that may severely degrade the reconstructed image. Therefore, error resilience is an essential problem in video communications. In this paper, we propose novel temporal error concealment techniques for recovering lost or erroneously received macroblock (MB). To reduce the computational complexity, the proposed method adaptively determines the search range for each lost MB to find best matched block in the previous frame. And the original corrupted MB split into for $8{\times}8$ sub-MBs, and estimates motion vector (MV) of each sub-MB using its boundary information. Then the estimated MVs are utilized to reconstruct the damaged MB. In simulation results, the proposed method shows better performance than conventional methods in both aspects of PSNR.