• Proceedings of the Korean Society of Computer Information Conference
• /
• 2013.01a
• /
• pp.229-230
• /
• 2013

본 논문에서는 입체영상을 분석하여 3차원의 영상 내에 나타나는 문자 영역을 효과적으로 분리하는 알고리즘을 제안한다. 제안된 알고리즘은 먼저 입력된 영상에서 질감 특징을 이용해 문자영역이 존재하는 후보 영역을 분할하고, 후보 문자영역 중에서 문자열만을 형성하는 영역을 추출한다. 그런 다음, 지역화된 문자영역을 문자와 배경으로 분리하며, 거리 특징을 활용하여 추출된 문자영역이 비 문자영역을 포함하지 않고 문자영역만을 포함하고 있는지를 최종적으로 검증한다. 실험에서는 제안된 방법을 여러 가지 영상에 적용하여 테스트 해 보았으며, 제안된 방법이 기존의 방법에 비해 보다 정확하게 문자영역을 추출함을 확인하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.133-134
• /
• 2014

Conventional stereo matching algorithms fail when they deal with anaglyph image as its input because anaglyph image does not have similar intensity on both view images. To ameliorate such problems, we propose a robust method to obtain accurate disparity maps. The novel Absolute Adaptive Normalized Cross Correlation (AANCC) for anaglyph data cost is introduced in this paper. Then, it is followed by occlusion detection and segmentation-based plane fitting to achieve accurate depth map acquisition. Experimental results confirm that the proposed anaglyph data cost is robust and gives accurate disparity maps.

• Proceedings of the IEEK Conference
• /
• 2003.11a
• /
• pp.393-396
• /
• 2003

This paper proposes a method to speed up the document image binarization using a water flow model. The proposed method extracts the region of interest (ROI) around characters from a document image and restricts pouring water onto a 3-dimensional terrain surface of an image only within the ROI. The amount of water to be filled into a local valley is determined automatically depending on its depth and slope. Then, the proposed method accumulates weighted water not only on the locally lowest position but also on its neighbors. Finally, the depth of each pond is adaptively thresholded for robust character segmentation. Experimental results on real document images shows that the proposed method has attained good binarization performance as well as remarkably reduced processing time compared with that of the existing method based on a water flow model.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.317-318
• /
• 2006

There are lot of noises and errors in depth map image which is gotten by using a stereo camera. These errors are caused by mismatching of the corresponding points which occur in texture-less region of input images of stereo camera or occlusions. In this paper, we use a method which is able to get rid of the noises through post processing and reduce the errors of disparity values which are caused by the mismatching in the texture-less region of input images through the correlation between the depth map images and the input images. Then we propose a novel method which segments the object by using the improved disparity map images and projections.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.1C
• /
• pp.53-61
• /
• 2008

In this paper, we propose an algorithm for generation of multi-view video from conventional 2 dimensional video. Color and motion information of an object are used for segmentation and from the segmented objects, multi-view video is generated. Especially, color information is used to extract the boundary of an object that is barely extracted by using motion information. To classify the homogeneous regions with color, luminance and chrominance components are used. A pixel-based motion estimation with a measurement window is also performed to obtain motion information. Then, we combine the results from motion estimation and color segmentation and consequently we obtain a depth information by assigning motion intensity value to each segmented region. Finally, we generate multi-view video by applying rotation transformation method to 2 dimensional input images and the obtained depth information in each object. The experimental results show that the proposed algorithm outperforms comparing with conventional conversion methods.

• Journal of Broadcast Engineering
• /
• v.17 no.6
• /
• pp.945-953
• /
• 2012

A depth map can be obtained by projecting/capturing patterns of stripes using a projector-camera system and analyzing the geometric relationship between the projected patterns and the captured patterns. This is usually called structured light technique. In this paper, we propose a new multi-threading scheme for accelerating a conventional structured light technique. On CPUs and GPUs, multi-threading can be implemented by using OpenMP and CUDA, respectively. However, the problem is that their performance changes according to the computational conditions of partial processes of a structured light technique. In other words, OpenMP (using multiple CPUs) outperformed CUDA (using multiple GPUs) in partial processes such as pattern decoding and depth estimation. In contrast, CUDA outperformed OpenMP in partial processes such as rectification and pattern segmentation. Therefore, we carefully analyze the computational conditions where each outperforms the other and do use the better one in the related conditions. As a result, the proposed method can estimate a depth map in a speed of over 25 fps on $1280{\times}800$ images.

• The KIPS Transactions:PartB
• /
• v.12B no.3 s.99
• /
• pp.247-256
• /
• 2005

One of the classic research problems in computer vision is that of stereo, i.e., the reconstruction of three dimensional shape from two or more images. This paper deals with the problem of extracting depth information of non-rigid dynamic 3D scenes from general 2D video sequences taken by monocular camera, such as movies, documentaries, and dramas. Depth of the blocks are extracted from the resultant block motions throughout following two steps: (i) calculation of global parameters concerned with camera translations and focal length using the locations of blocks and their motions, (ii) calculation of each block depth relative to average image depth using the global parameters and the location of the block and its motion, Both singular and non-singular cases are experimented with various video sequences. The resultant relative depths and ego-motion object shapes are virtually identical to human vision.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.3
• /
• pp.389-395
• /
• 2022

Because of improvement of deep learning techniques, deep learning using computer vision such as classification, detection and segmentation has also been used widely at many fields. Expecially, automatic driving is one of the major fields that applies computer vision systems. Also there are a lot of works and researches to combine multiple tasks in a single network. In this study, we propose the network that predicts the individual depth of pedestrians and vehicles. Proposed model is constructed based on YOLOv3 for object detection and Monodepth for depth estimation, and it process object detection and depth estimation consequently using encoder and decoder based on hard parameter sharing. We also used attention module to improve the accuracy of both object detection and depth estimation. Depth is predicted with monocular image, and is trained using self-supervised training method.

• ETRI Journal
• /
• v.29 no.3
• /
• pp.353-362
• /
• 2007

In this paper, we propose a novel unsupervised video object extraction algorithm for individual images or image sequences with low depth of field (DOF). Low DOF is a popular photographic technique which enables the representation of the photographer's intention by giving a clear focus only on an object of interest (OOI). We first describe a fast and efficient scheme for extracting OOIs from individual low-DOF images and then extend it to deal with image sequences with low DOF in the next part. The basic algorithm unfolds into three modules. In the first module, a higher-order statistics map, which represents the spatial distribution of the high-frequency components, is obtained from an input low-DOF image. The second module locates the block-based OOI for further processing. Using the block-based OOI, the final OOI is obtained with pixel-level accuracy. We also present an algorithm to extend the extraction scheme to image sequences with low DOF. The proposed system does not require any user assistance to determine the initial OOI. This is possible due to the use of low-DOF images. The experimental results indicate that the proposed algorithm can serve as an effective tool for applications, such as 2D to 3D and photo-realistic video scene generation.

• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.9
• /
• pp.745-752
• /
• 2016

We suggest an efficient Simultaneous Localization and 3D Polygon Map Building (SLAM) method with Kinect depth sensor for mobile robots in indoor environments. In this method, Kinect depth data is separated into row planes so that scan line segments are on each row plane. After grouping all scan line segments from all row planes into line groups, a set of 3D Scan polygons are fitted from each line group. A map matching algorithm then figures out pairs of scan polygons and existing map polygons in 3D, and localization is performed to record correct pose of the mobile robot. For 3D map-building, each 3D map polygon is created or updated by merging each matched 3D scan polygon, which considers scan and map edges efficiently. The validity of the proposed 3D SLAM algorithm is revealed via experiments.