• Journal of Korea Multimedia Society
• /
• v.16 no.6
• /
• pp.764-781
• /
• 2013

In this paper, we propose a highly precise vehicle detection method with low false alarm using billboard sweep stereo matching and multi-stage hypothesis generation. First, we capture stereo images from cameras established in front of the vehicle and obtain the disparity map in which the regions of ground plane or background are removed using billboard sweep stereo matching algorithm. And then, we perform the vehicle detection and tracking on the labeled disparity map. The vehicle detection and tracking consists of three steps. In the learning step, the SVM(support vector machine) classifier is obtained using the features extracted from the gabor filter. The second step is the vehicle detection which performs the sobel edge detection in the image of the left camera and extracts candidates of the vehicle using edge image and billboard sweep stereo disparity map. The final step is the vehicle tracking using template matching in the next frame. Removal process of the tracking regions improves the system performance in the candidate region of the vehicle on the succeeding frames.

• Journal of Broadcast Engineering
• /
• v.25 no.5
• /
• pp.685-697
• /
• 2020

Immersive media videos, such as panorama and 360-degree videos, must provide a sense of realism as if the user visited the space in the video, so they should be able to represent the reality of the real world. However, in panorama and 360-degree videos, objects appear to overlap or disappear due to parallax between cameras, and such parallax distortion may interfere with immersion of the user's content. Accordingly, although many video stitching algorithms have been proposed to overcome parallax distortion, parallax distortion still occurs due to the low performance of the Object detection module and limitations of the Seam generation method. Therefore, this paper analyzes the limitations of the existing video stitching technology and proposes a method for detecting and correcting parallax distortion of video stitching using the LDPM (Local Differential Pixel Mean) image evaluation method that overcomes the limitations of the video stitching technique.

• Proceedings of the Korean Association of Geographic Inforamtion Studies Conference
• /
• 2004.03a
• /
• pp.511-516
• /
• 2004

수치표고모델, 정사영상과 같은 공간영상정보를 구축하기 위해서는 입체영상을 이동한 영상정합(image matching)의 과정이 필수적이며, 단영상 또는 스테레오 영상을 이용하여 대상물의 3차원 정보를 재구성하고 복원하는 기술은 사진측량 및 컴퓨터 비전 분야의 주요 연구 중의 하나이다. 본 연구에서는 화소값의 유사성과 상호관계성을 고려하는 MRF 모델을 이용하여 영상정합을 수행하였다. MRF 모델은 공간분석이나 물리적 현상의 전후관계(contextural dependencies)의 분석을 위한 확률이론의 한 분야로 다양한 공간정보를 통합할 수 있는 방법을 제공한다. 본 연구에서는 기준영상의 화소에 시차를 할당하는 접근 방법으로 확률모델의 일종인 마르코프 랜덤필드(MRF)모델에 기반한 영상정합기법을 제안하였고, 공간내 화소의 상호관계를 고려해주므로 대상물의 경계부분에서의 매칭 정확도를 향상시켰다. 영상정합문제에서의 MRF 기본가정은 영상 내 특정화소의 시차는 그 주위화소의 시차에 의한 부분정보에 따라 결정이 가능하다는 것이다. 깁스분포(gibbs distribution)를 사용하여 사후(posteriori) 확률값을 유도해내고, 이를 최대사후확률(MAP: Maximum a Posteriori)추정법을 이용하여 에너지함수를 생성하였다. 생성된 에너지함수의 최적화(Optimization)를 위하여 본 연구에서는 전역최적화기법인 multiway cut 기법을 사용하여 영상정합에 있어 에너지함수를 최소로 하는 이미지화소에 대한 시차레이블을 구하여 영상정합을 수행하였다.

• Journal of the Institute of Convergence Signal Processing
• /
• v.2 no.2
• /
• pp.20-27
• /
• 2001

3D image coding and transmission technologies are essential for transmitting 3D image information through the limited transmission channel, and very important in the area of information and telecommunication services using 3D image media. In this paper, we propose the 3D image coding and transmission methods welch use disparity information and MPEG-2's scalability. Left and Fight image sequences are coded in base layer and enhancement layer, respectively. The enhancement layer contains the disparity and prediction informations. In order to reduce searching time for disparity information, we propose a method in which the disparity information of previous image paid is used in the next image pair, and we adopt the bit rate control method proposed in MPEG-2 TM6. An experimental result showed that the proposed searching method considerably reduced the entire encoding time and reached the desired bit rates.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.48 no.6
• /
• pp.77-87
• /
• 2011

In this paper, we propose a disparity correction technique to reduce the inherent visual discomfort while watching stereoscopic videos. The visual discomfort must be solved for commercial 3D display systems to provide natural stereoscopic videos to human eyes. The proposed disparity correction technique consists of horizontal and vertical disparity corrections. The horizontal disparity correction is implemented by controlling the depth budget of stereoscopic video using the geometric relations of a stereoscopic camera system. In addition, the vertical disparity correction is implemented by using a feature-based stereo matching algorithm. Conventional vertical disparity corrections have been done by only using camera calibration parameters, which still cause systematic errors in vertical disparities. In this paper, we minimize the vertical disparity as small as possible by using a feature-based correction algorithm. Through the comparisons of conventional feature-based correction algorithms, we analyze the performance of the proposed technique.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.4 s.304
• /
• pp.13-20
• /
• 2005

In this paper, we propose new smoothing filters, i.e., occluding patterns that can accurately correct disparities of occluded areas in the estimated disparity map. An image is composed of several layers and each layer presents similar disparity. Furthermore, the distribution of the estimated disparities has a specific direction around the boundary of the occlusion, and this distribution presents the different direction with respect to the left- and the right-based disparity map. However, typical smoothing filters, such as mean filter and median filter, did not take into account those characteristic. So, they can decrease some error, but they cannot guarantee the accuracy of the corrected disparity. On the contrary, occluding patterns can accurately correct disparities of occluded areas because they consider both the characteristic that occlusion occurs and the characteristic that disparities of the occlusion are ranged, from estimated disparity maps with respect to the left and the right images. We made experiments on occluding patterns with some real stereo image set, and as a result, we can correct disparities of occluded areas more accurately than typical smoothing filters did.

• The KIPS Transactions:PartB
• /
• v.19B no.2
• /
• pp.113-122
• /
• 2012

This paper presents a new stereo matching technique which is based on the matching of feature windows and feature links. The proposed method uses the FAST feature detector to find image features in stereo images and determines the correspondences of the detected features in the stereo images. We define a feature window which is an image region containing several image features. The proposed technique consists of two matching steps. First, a feature window is defined in a standard image and its correspondence is found in a reference image. Second, the corresponding features between the matched windows are determined by using the feature link technique. If there is no correspondence for an image feature in the standard image, it's disparity is interpolated by neighboring feature sets. We evaluate the accuracy of the proposed technique by comparing our results with the ground truth of in a stereo image database. We also compare the matching accuracy and computation time with two conventional feature-based stereo matching techniques.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2005.05a
• /
• pp.851-854
• /
• 2005

본 논문에서는 영상 내 객체들의 운동시차를 이용하여 실시간으로 2차원 영상을 입체 영상으로 변환하는 알고리즘을 제안하였다. 제안한 입체 영상 변환 알고리즘은 영상 내 객체 분할, 운동시차를 이용한 운동 객체 검출, 시차 처리를 이용하여 입체 영상으로 변환한다. 입체 영상 변환 알고리즘의 성능 평가를 위해 좌안과 우안 영상의 절대 차이 영상을 이용하여 기존의 MTD 방식과 시뮬레이션 결과를 비교, 분석하였다. 제안한 알고리즘은 영상의 운동 방향 및 속도에 상관없이 다양한 영상원에 대해 실시간 입체 영상 변환이 가능하다. 입체 영상 변환 장치는 VHDL로 설계하였으며, FPGA 테스트 보드 구현을 통해 성능 및 기능을 검증하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06c
• /
• pp.521-524
• /
• 2010

영상을 입체적으로 보는 방법에는 안경을 이용한 편광 방식과 시분할 방식, 안경이 필요 없는 패럴랙스 베리어, 렌티큘러, 다시점 방식 그리고 완전 3차원 방법들이 있다. 그 중에서도 안경을 이용하여 가장 쉽게 제작이 가능하고, 비용이 저렴한 애너그리프(Anaglyph) 방법이 흔히 사용된다. 애너그리프 영상을 생성하는 방법에도 여러 가지가 존재하고 최근까지 눈에 피로를 적게 하면서 입체감을 최대한 표현할 수 있는 영상을 생성하기 위한 연구가 진행되고 있다. 본 논문에서는 이러한 조건을 만족시키기 위하여 새로운 애너그리프 영상 생성 방법을 제안한다. 본 논문에서 제안하는 방법은 깊이 맵과 양안시차를 계산하여 가장 입체감 있는 애너그리프 영상을 생성 하는 방법이다. 깊이 맵의 계산을 통해 얻은 변이 정보와 실험적으로 계산한 사람 눈의 양안시차를 좌측 영상에 적용한다. 좌측 영상과 우측 영상을 최적화된 색상혼합방법을 사용하여 합성하면 제안한 방식의 애너그리프 영상이 생성된다. 본 논문에서 제안하는 방법은 기존의 초점을 고려하지 않는 애너그리프 방식의 문제점을 깊이 맵을 이용하여 해결할 수 있고 또한 양안시차를 고려하고 최적화된 색상혼합을 사용하기 때문에 눈에 피로가 적어진다.

• Journal of Broadcast Engineering
• /
• v.17 no.3
• /
• pp.491-502
• /
• 2012

Flat panel-based parallax barrier or lenticular based 3D display devices that have been developed recently are designed to feel depth. In order to see a three-dimensional(3D) image by the display device, a multi-view image displayed on the flat panel must be regenerated from images of multi-views using a subsampling method. Previous subsampling methods are focused on reducing crosstalk. In this paper, we focus on a misalignment that is occurred on manufacture process of slanted parallax barrier based autostereoscopic display device. Therefore, we propose a interleaving method that considers tilt of slanted parallax barrier, aperture size, and distance between an autostereoscopic display device and a viewer to see a 3D image regardless of a viewer position.