• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.89-92
• /
• 1996

In this paper, we developed visual tracking algorithm using double active bar. The active bar model to represent the object can reduce the search space of energy surface and better performance than those of snake model. However, the contour will not find global equilibrium when driving force caused by image may be weak. To overcome this problem. Double active bar is proposed for finding the global minimum point without any dependence on initialization. To achieve the goal, an deformable model with two initial contours in attempted to search for a global minimum within two specific initial contours. This approach improve the performance of finding the contour of target. To evaluate the performance, some experiments are executed. We can achieved the good result for tracking a object on noisy image.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1999.06a
• /
• pp.35.2-42
• /
• 1999

When we see a binary black and white image, it appears. to our visual sense, to become clearer because of a pseudo halftone. This is because when we don't notice of all the details of the image, instead we see from a more global standpoint. We presented a theory and experimental results for layered model which extended external world, retina and brain of the two dimensional visual model. This paper propose the objective evaluation to coincide with the subjective evaluation of the human in ordered to evaluate the relative merits of the various modulation method. We obtain experimentally the restored images and the measure of approximation of the density four division and ordered dither method. The measure of approximation present s the objective evaluation scale to coincide with the subjective evaluation of the two modulation images.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.3
• /
• pp.853-873
• /
• 2021

In recent years, deep reinforcement learning (DRL) models are enjoying great interest as their success in a variety of challenging tasks. Deep Q-Network (DQN) is a widely used deep reinforcement learning model, which trains an intelligent agent that executes optimal actions while interacting with an environment. This model is well known for its ability to surpass skilled human players across many Atari 2600 games. Although DQN has achieved excellent performance in practice, there lacks a clear understanding of why the model works. In this paper, we present a visual analytics system for understanding deep Q-network in a non-blind matter. Based on the stored data generated from the training and testing process, four coordinated views are designed to expose the internal execution mechanism of DQN from different perspectives. We report the system performance and demonstrate its effectiveness through two case studies. By using our system, users can learn the relationship between states and Q-values, the function of convolutional layers, the strategies learned by DQN and the rationality of decisions made by the agent.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.20 no.3
• /
• pp.6-12
• /
• 1983

The visual evoked potentials were measured using 19 electrodes attached to the scalp in 5 normal and 4 abnormal subjects during visual stimulation and these data were sampled for computer processing with 500 Hz sampling frequency. The center of potential distributions and its time-dependent locus were estimated from these potential distributions using weighting matrix which was determined by the electrodes' position coordinates. In normal subjects these estimated electrical signals were shown to propagate from the frontal lobe to the occipital lobe of the cortex following the known visual pathway. In abnormal subjects, there were significant differences in these estimated propagation pathway. The relationships among this model, the point source model and the dipole source model were analyzed.

• Korean Journal of Cognitive Science
• /
• v.8 no.4
• /
• pp.63-75
• /
• 1997

Changes in visual sensitivity were investigated as a function of distance from the locus of attention. While a subject was fixating at a point on a frontal plane, one of the two attention inducing points placed horizontally and symmetrically 4。 apart from it was blinked and a target immediately followed at a location around the blinking dot. The subject's task was to decide and report whether the target was present or abscent. The rate of detection was the highest at the immediate vicinity of the locus of attention and decreased gradually as a function of the distance from it. Results of the experiments support the gradient model of attention-induced changes in visual sensitivity.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.4A
• /
• pp.371-382
• /
• 2002

A camera model with optical axes parallel has been widely used for stereo vision applications. A pair of input ages are obtained from a convergent-looking stereo camera model based on the human visual system in this per, and each image is divided into quadrant regions with respect to the fixation point. The reasoning of quadrant partitions is based on the human visual system and is proven by a geometrical method. Image patches : constructed from the right and left stereo images. A modified cepstrum filter is applied to the patches and disparity vectors are determined by peak detection algorithm. The three-dimensional information for synthetic ages is obtained from the measured disparity and the convergent stereo camera model. It is shown that the experimental results of the proposed method for various stereo images are accurate around the fixation point like the human visual system.

• International Journal of Internet, Broadcasting and Communication
• /
• v.14 no.1
• /
• pp.129-135
• /
• 2022

Real human digital models appear more and more frequently in VR/AR application scenarios, in which real-time markerless face capture animation of personalized virtual human faces is an important research topic. The traditional way to achieve personalized real human facial animation requires multiple mature animation staff, and in practice, the complex process and difficult technology may bring obstacles to inexperienced users. This paper proposes a new process to solve this kind of work, which has the advantages of low cost and less time than the traditional production method. For the personalized real human face model obtained by 3D reconstruction technology, first, use R3ds Wrap to topology the model, then use Avatary to make 52 Blend-Shape model files suitable for AR-Kit, and finally realize real-time markerless face capture 3D real human on the UE4 platform facial motion capture, this study makes rational use of the advantages of software and proposes a more efficient workflow for real-time markerless facial motion capture of personalized 3D real human models, The process ideas proposed in this paper can be helpful for other scholars who study this kind of work.

• Journal of Digital Contents Society
• /
• v.18 no.7
• /
• pp.1313-1321
• /
• 2017

In the future society where ICT-based digital convergence creates new value, collaborative skills among experts in various fields and SW based problem solving ability is more emphasized. Non-SW specialists are required to have SW based communication skills to effectively collaborate with SW experts to solve their problems. Therefore, SW programming curriculum for non-major college students should be different from the existing programming education for SW-majors aiming at a high level of coding ability. It is also known that diagram-based visual representation is helpful for productive communication and collaboration. In this study, we defined the SW education objectives for the non-majors as cultivating the visual programming ability for SW based problem solving. In order to accomplish this, we explored SW programming curriculum and content development model for non-majors focusing on visual representation of SW solutions. The results of this paper will help to provide appropriate SW learning model for non-majors and to cultivate practical SW capabilities.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.53 no.1
• /
• pp.50-56
• /
• 2020

Determination of the precise size at sexual maturity is very important for science-based stock assessment and fisheries resource management. In this study, two different models, (1) a visual method and (2) a gonadosomatic index (GSI) method, were employed to estimate length at sexual maturity of the small yellow croaker Larimichthys polyactis in the Yellow Sea of Korea. The visual method is a common qualitative method using visual gonadal identification. Conversely, the GSI method is a quantitative method using the GSI, which can be easily and precisely collected. We compared results from these methods to determine the best approach, and to examine the practicality of the GSI method. Logistic regression of the maturity ogive was conducted using a general linear model (GLM) with the R statistics program. Also, the bootstrapped 95% confidence intervals of all estimates were calculated. The best-fit model was the visual method (R_Mc²=0.805, AUC=0.989, L₅₀=15.1). Among models using the GSI method, the model computing GSI_ref=0.94 was the best-fit model (R_Mc²=0.792, AUC=0.989, L₅₀=15.2). There was no significant difference between the two models, evidencing the effectiveness and accuracy of the GSI method.

• Journal of the Korea Computer Graphics Society
• /
• v.23 no.4
• /
• pp.21-29
• /
• 2017

Previous visual analytics researches has focused on reducing the uncertainty of predicted results using a variety of interactive visual data exploration techniques. The main purpose of the interactive search technique is to reduce the quality difference of the predicted results according to the level of the decision maker by understanding the relationship between the variables and choosing the appropriate model to predict the unknown variables. However, it is difficult to create a predictive model which forecast time series data whose overall trends is unknown such as youth physical growth data. In this paper, we pro pose a novel predictive analysis technique to forecast the physical growth value in small pieces of time series data with un certain trends. This model estimates the distribution of data at a particular point in time. We also propose a visual analytics system that minimizes the possible uncertainties in predictive modeling process.