• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 한국방송∙미디어공학회 2021년도 추계학술대회
• /
• pp.41-44
• /
• 2021

In this paper, we propose an efficient reference-based compression artifact reduction network for 360-degree images in an equi-rectangular projection (ERP) domain. In our insight, conventional image restoration methods cannot be applied straightforwardly to 360-degree images due to the spherical distortion. To address this problem, we propose an adaptive disparity estimator using a deformable convolution to exploit correlation among 360-degree images. With the help of the proposed convolution, the disparity estimator establishes the spatial correspondence successfully between the ERPs and extract matched textures to be used for image restoration. The experimental results demonstrate that the proposed algorithm provides reliable high-quality textures from the reference and improves the quality of the restored image as compared to the state-of-the-art single image restoration methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제18권3호
• /
• pp.653-669
• /
• 2024

Dynamic scene deblurring is a complex computer vision problem owing to its difficulty to model mathematically. In this paper, we present a novel approach for image deblurring with the help of the sharp reference image, which utilizes the reference image for high-quality and high-frequency detail results. To better utilize the clear reference image, we develop an encoder-decoder network and two novel modules are designed to guide the network for better image restoration. The proposed Reference Extraction and Aggregation Module can effectively establish the correspondence between blurry image and reference image and explore the most relevant features for better blur removal and the proposed Spatial Feature Fusion Module enables the encoder to perceive blur information at different spatial scales. In the final, the multi-scale feature maps from the encoder and cascaded Reference Extraction and Aggregation Modules are integrated into the decoder for a global fusion and representation. Extensive quantitative and qualitative experimental results from the different benchmarks show the effectiveness of our proposed method.

• Korean Journal of Cognitive Science
• /
• 제12권4호
• /
• pp.49-56
• /
• 2001

When multiple features are presented in the image the computational models for stereopsis select the most activated matches through the excitatory and inhibitory interactions among all possible matches. Using the random-dot stereogram with two surfaces. we investigate whether human binocular mechanism selects the most activated matches. The dots consisting a surface lying in a fixation plane were selected randomly while each of the dots consisting the other surface was paired with each of the original dots in the following manner. After finding the position of each dots in the original random pattern we placed an additional dot to the left and to the right of the original position in each of the left and right image of a stereogram respectively. The luminance of additional dots was varied while that of the original random dots was fixed so that the hypothetical matches presumably could be activated differently. Across the luminance condition the depth of each surface was measured to examine whether matches to be selected were changed depending on the activation level of possible matches. When the luminance of two patterns was within 30% of one another observers perceived an opaque surface. Beyond this value two transparent surfaces were seen with the magnitude of perceived depth varying with relative luminance of two patterns. When original pattern was brighter one additional surface was perceived at the depth corresponding to the disparity of original pattern. When original dot was dimmer. however the depth of an additional surface corresponded to the disparity of newly introduced pattern. These results suggest that there are dynamic interactions within the matching process whereby highly activated matches inhibit weaker one.

• Journal of the Institute of Electronics and Information Engineers
• /
• 제50권1호
• /
• pp.75-85
• /
• 2013

To acquire depth information using stereo vision, it is required to find correspondence points between stereo image pair. Conventional stereo vision systems usually use two cameras to get disparity data. Therefore, conventional stereo matching methods cannot resolve the tradeoff problem between accuracy and precision with respect to the length of baseline. Besides, belief propagation method, which is being used recently, has a problem that matching performance is dependent on the fixed weight parameter ${\lambda}$. In this paper, we propose a modified belief propagation stereo matching technique based on multi-baseline stereo vision to solve the tradeoff problem. The proposed method calculates EMAD(extended mean of absolute differences) as local evidence. And proposed method decides weight parameter ${\lambda}$ adaptively to local texture information. The proposed method shows higher initial matching performance than conventional methods and reached optimum solution in less iteration. The matching performance is increased about 4.85 dB in PSNR.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• 제13권6호
• /
• pp.1194-1199
• /
• 2009

Stereo vision system is useful method for inferring 3D depth information from two or more images. So it has been the focus of attention in this field for a long time. Stereo matching is the process of finding correspondence points in two or more images. A central problem in a stereo matching is that it is difficult to satisfy both the computation time problem and accuracy at the same time. To resolve this problem, we proposed a new stereo matching technique using weighted cost aggregation. To begin with, we extract the weight in given stereo images based on features. We compute the costs of the pixels in a given window using correlation of weighted color and brightness information. Then, we match pixels in a given window between the reference and target images of a stereo pair. To demonstrate the effectiveness of the algorithm, we provide experimental data from several synthetic and real scenes. The experimental results show the improved accuracy of the proposed method.

• Journal of Digital Contents Society
• /
• 제5권3호
• /
• pp.163-170
• /
• 2004

Ubiquitous is an agriculture revolution, the Industrial Revolution, the fourth revolution to equal an information revolution are Ubiquitous. Historic agriculture was revolutionary, but the Industrial Revolution was a revolution of the physics space that was a base of a mankind civilization, and the informatization revolution that Cyber Space saw a prime while world wide web service was expended was a revolution of cyber space. Be charmed with this, and, as for the Ubiquitous revolution, it is achieved a physics space and an intelligent union of cyber space in the space where an off-line space was integrated with compunction on-line. It is combined with a life space naturally, and Ubiquitous technology. Also, I presented the plan that derivation tried a problem about Ubiquitous, and can solve an each problem. Specially, infringement about personal privacy very has a lot of possession to occur in the world that radio communication between all appliances is possible. It is Ubiquitous by becoming the data which is useful to thorough preparation about the Ubiquitous world to come with presenting policy a little technical solution plan about this to the future Republic of Korea. many company has been cracked by crackers information security and everyday new computer virus come out.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 한국해양정보통신학회 2009년도 춘계학술대회
• /
• pp.396-399
• /
• 2009

Stereo vision system is useful method for inferring 3D depth information from two or more images. So it has been the focus of attention in this field for a long time. Stereo matching is the process of finding correspondence points in two or more images. A central problem in a stereo matching is that it is difficult to satisfy both the computation time problem and accuracy at the same time. To resolve this problem, we proposed a new stereo matching technique using weighted cost aggregation. To begin with, we extract the weight in given stereo images based on features. We compute the costs of the pixels in a given window using correlation of weighted color, brightness and distance information. Then, we match pixels in a given window between the reference and target images of a stereo pair. To demonstrate the effectiveness of the algorithm, we provide experimental data from several synthetic and real scenes. The experimental results show the improved accuracy of the proposed method.

• Journal of the Korean Geographical Society
• /
• 제29권1호
• /
• pp.65-83
• /
• 1994

Sevaral problems of administrative area sysem in Korea have been brought up for a long time. Because its frame has remained since Chosun and Japanese colonial period in spite of changing local administrative environment in accordance with rapid industrialization and urbanization. Recent reform of city (Shi)- county (Gun) integration is derived from this argument. But problems which permeate deeply overall system cannot be solved by partial reorganization of Shi-Gun. They may be rationalized only through the reform of the whole system. The aims of this study are to analyze problems of administrative area system entirelr and to discuss the direction of its reform from that point of view. Major problems of administrative area system are summed up into the followings. Firstly, it is found that administrative hierarchies are too many levels. Contemporary administrative hierarchical structure is 4 levels: regional autonomous government (Tukpyolshi, Jik'halshi, Do), local autonomous government (Shi, Gun), two leveis of auxiliary administrative area (Up, Myun and Ri). These hierarchies were established in late period of Chosun which transportation was undeveloped and residential activity space was confined. But today developing transportion and expanding sphere of life don't need administrative hierarchical structurl with many levels. Besides developing administrative technology reduces administrative space by degrees. Many levels of contemporary administrative hierarchical structure are main factor of administrative inefficency, discording with settlement system. Second problem is that Tukpyolshi and Jik'halshi - cities under direct control of the central government as metropolitan area - underbounded cities. Underbounded city discomforts residential life and increases external elects of local pulic services. Especially this problem is Seoul, Pusan and Daegu. Third problem is that Do-areas are mostly two larger in integrating into single sphere of life. In fact each of them consistes of two or three sphere of life. Fourth Problem is metropolitan government system that central city is seperated from complementary area, i.e. Do. It brings about weakening the economic force of Do. Fifth problem is that several cities divided single sphere of life. It is main factor of finantial inefficency and facing difficult regional administration. Finally necessity of rural parish (Myun.) is diminished gradually with higher order center oriented activty of rural residents. First of all administrative area system should corresponds with substantial sphere of life in order to solve these problems. Followings are some key directions this study proposes on the reform of administrative area system from that standpoint. 1. Principles of reorgnization -- integration of central dty with complementary area. -- correspondence of administrative hierarchical structure with settlement system. -- correspondence of boundary of administrative area with sphere of life. 2. Reform strategy -- Jik'halshi is integrated with Do and is under the contol of Do. -- Small Seoul shi (city) which have special functions as captal is demarcated in Seoul tukpyolshi and 22 autonomous distrcts of Seoul tukpyolshi is integrated into 3-4 cities. -- Neighboring cities (Shies) in single sphere of life are intrgrated into single city (Shj). -- Myun and Ri are abolished in rural region and new unit of local administrative area on the basis of lowest order sphere of life into which 3-4 Ries are integrated replaces them.

• Korean Journal of Computational Design and Engineering
• /
• 제16권4호
• /
• pp.249-259
• /
• 2011

In this paper we present our work on the parameterized construction of virtual drivers' reach motion to seat belt, by using motion capture data. A user can generate a new reach motion by controlling a number of parameters. We approach the problem by using multiple sets of example reach motions and learning the relation between the labeling parameters and the motion data. The work is composed of three tasks. First, we construct a motion database using multiple sets of labeled motion clips obtained by using a motion capture device. This involves removing the redundancy of each motion clip by using PCA (Principal Component Analysis), and establishing temporal correspondence among different motion clips by automatic segmentation and piecewise time warping of each clip. Next, we compute motion blending functions by learning the relation between labeling parameters (age, hip base point (HBP), and height) and the motion parameters as represented by a set of PC coefficients. During runtime, on-line motion synthesis is accomplished by evaluating the motion blending function from the user-supplied control parameters.

• The KIPS Transactions:PartB
• /
• 제17B권1호
• /
• pp.63-68
• /
• 2010

In this paper, we propose a multiresolution stereo matching method based on genetic algorithm using edge information. The proposed approach considers the matching environment as an optimization problem and finds the solution by using a genetic algorithm. A cost function composes of certain constraints which are commonly used in stereo matching. We defines the structure of chromosomes using edge pixel information of reference image of stereo pair. To increase the efficiency of process, we apply image pyramid method to stereo matching and calculate the initial disparity map at the coarsest resolution. Then initial disparity map is propagated to the next finer resolution, interpolated and performed disparity refinement. We valid our approach not only reduce the search time for correspondence but alse ensure the validity of matching.