• Journal of KIISE
• /
• v.44 no.2
• /
• pp.186-194
• /
• 2017

Depth-image-based rendering is a technique that can be applied in a variety of 3D-display systems. It generates the images that have been captured from virtual viewpoints by using a depth map. However, disoccluded hole-filling problems remain a challenging issue, as a newly exposed area appears in the virtual view. Image inpainting is a popular approach for the filling of the hole region. This paper presents a robust hole-filling method that reduces the error and generates a high quality-virtual view. First, the adaptive-patch size is decided using the color and depth information. Also, a partial filling method for which the patch similarity is used is proposed. These efforts reduce the error occurrence and the propagation. The experiment results show that the proposed method synthesizes the virtual view with a higher visual comfort compared with the existing methods.

• The Journal of Information Systems
• /
• v.31 no.4
• /
• pp.145-160
• /
• 2022

Purpose The purpose of this study is to develop a virtual try-on deep learning model that can efficiently learn front and back clothes images. It is expected that the application of virtual try-on clothing service in the fashion and textile industry field will be vitalization. Design/methodology/approach The data used in this study used 232,355 clothes and product images. The image data input to the model is divided into 5 categories: original clothing image and wearer image, clothing segmentation, wearer's body Densepose heatmap, wearer's clothing-agnosting. We advanced the HR-VITON model in the way of Mixed-Precison, Gradient Accumulation, and sharing model weights. Findings As a result of this study, we demonstrated that the weight-shared MP-GA HR-VITON model can efficiently learn front and back fashion images. As a result, this proposed model quantitatively improves the quality of the generated image compared to the existing technique, and natural fitting is possible in both front and back images. SSIM was 0.8385 and 0.9204 in CP-VTON and the proposed model, LPIPS 0.2133 and 0.0642, FID 74.5421 and 11.8463, and KID 0.064 and 0.006. Using the deep learning model of this study, it is possible to naturally fit one color clothes, but when there are complex pictures and logos as shown in <Figure 6>, an unnatural pattern occurred in the generated image. If it is advanced based on the transformer, this problem may also be improved.

• Journal of the Korean Society of Earth Science Education
• /
• v.8 no.2
• /
• pp.193-205
• /
• 2015

In this study, a Virtual geological Field Trip(VFT) learning program using 3D panorama virtual reality techniques was developed to learn about the Gongju city 7 area located in Chungcheongnam-do, Korea. The developed $360^{\circ}$ 3D VFT program can show every face of observational points and interact as zoom-in, zoom-out and image rotation. For the educational effects of the materials, it is provided with a compass, a protractor, enlarged images, pop-up windows, etc.. The program was applied to the class of 35 gifted students in middle school to investigate the effectiveness of the program. The results showed that positive responses of the students were 90% or more. When geological field trip problems like cost, safety, distance occur in geological learning procedure of middle school science, this VFT program can become as a supplementary learning material and a solution.

• International Journal of Internet, Broadcasting and Communication
• /
• v.11 no.2
• /
• pp.43-49
• /
• 2019

Nowadays, one-person media broadcasting has become popular, and with the progress of this popularity, multimedia techniques which can support such broadcasting are also becoming more and more advanced. One of the most emerging multimedia technique used in this field is the virtual reality technology which sets the one-person media broadcasting environment as a virtual reality environment. However, as such an environment requires instruments of high cost, it is not easy for normal individuals to constitute such environments. Therefore, in this paper we propose how to construct virtual reality-like panoramas with a multiple of smartphones. For this purpose, we designed a special rig which can hold firmly 8 smartphone cameras which have overlapping view of the environment such that panorama stitching becomes possible. To reduce the computation cost, we precomputed the homography matrices, and used 1-D pointer structures to store the computed coordinate values.

• Journal of the Korean Society of Visualization
• /
• v.1 no.1
• /
• pp.92-97
• /
• 2003

A new Stereoscopic PTV was developed using two CCD cameras, stereoscopic photogrammetry based on a 3D-PTV principle. Virtual images were produced for the benchmark test of the constructed Stereoscopic PTV technique. The arrangement of the two cameras was based on angular position. The calibration of cameras and the pair-matching of the three-dimensional velocity vectors were based on the Genetic Algorithm based 3D-PTV technique. The constructed Stereoscopic PTV technique was tested on the standard images of the impinged jet proposed by VSJ. The results obtained by the constructed system showed good agreements with the original data.

• Proceedings of the KSME Conference
• /
• 2001.11b
• /
• pp.394-399
• /
• 2001

A new stereoscopic PIV is introduced. The system works with CCD cameras, stereoscopic photogrammetry, and a 3D-PTV principle. Virtual images are produced for the construction of a benchmark testing tool of PIV techniques. The arrangement of the two cameras is based on angular position. The calibration of cameras and the pair-matching of the three-dimensional velocity vectors are based on 3D-PTV technique.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.38 no.5
• /
• pp.407-414
• /
• 2020

SVM (Support vector machine) analysis was performed after applying a deep learning technique based on an Inception-based model (GoogLeNet). The accuracy of automatic image classification was analyzed using an SVM with multiple virtual grid sizes. Six classes were selected from a standard land cover map. Cars were added as a separate item to increase the classification accuracy of roads. The virtual grid size was 2-5 m for natural areas, 5-10 m for traffic areas, and 10-15 m for building areas, based on the size of items and the resolution of input images. The results demonstrate that automatic classification accuracy can be increased by adopting an integrated approach that utilizes weighted virtual grid sizes for different classes.

• Journal of information and communication convergence engineering
• /
• v.15 no.4
• /
• pp.256-261
• /
• 2017

In this paper, we propose a modified smart pixel mapping (SPM) to visualize occluded three-dimensional (3D) objects in real image fields. In integral imaging, orthoscopic real 3D images cannot be displayed because of lenslets and the converging light field from elemental images. Thus, pseudoscopic-to-orthoscopic conversion which rotates each elemental image by 180 degree, has been proposed so that the orthoscopic virtual 3D image can be displayed. However, the orthoscopic real 3D image cannot be displayed. Hence, a conventional SPM that recaptures elemental images for the orthoscopic real 3D image using virtual pinhole array has been reported. However, it has a critical limitation in that the number of pixels for each elemental image is equal to the number of elemental images. Therefore, in this paper, we propose a modified SPM that can solve this critical limitation in a conventional SPM and can also visualize the occluded objects efficiently.

• Journal of KIISE:Computer Systems and Theory
• /
• v.29 no.8
• /
• pp.463-471
• /
• 2002

In order to implement virtual endoscopy, high-speed rendering algorithm that generates accurate perspective projection images and efficient collision detection method are essential. In this paper we propose an efficient virtual endoscopy system based on volume rendering technique. It is possible to skip over empty (transparent) space using the distance value produced in preprocessing time, and it does not deteriorate image quality since it is an extension of ray-casting. It also accelerates rendering speed with minimal loss of image quality by adjusting sampling interval along a ray according to direction of the ray. Using the distance information, we can simplify the collision detection of volumetric objects.

• Cartoon and Animation Studies
• /
• s.47
• /
• pp.1-29
• /
• 2017

360-degree virtual reality has been a technology that has been available for a long time and has been actively promoted worldwide in recent years due to development of devices such as HMD (Head Mounted Display) and development of hardware for controlling and executing images of virtual reality. The production of the 360 degree VR requires a different mode of production than the traditional video production, and the matters to be considered for the user have begun to appear. Since the virtual reality image is aimed at a platform that requires enthusiasm, presence and interaction, it is necessary to have a suitable cinematography. In VR, users can freely enjoy the world created by the director and have the advantage of being able to concentrate on his interests during playing the image. However, the director had to develope and install the device what the observer could concentrate on the narrative progression and images to be delivered. Among the various methods of transmitting images, the director can use the composition of the short. In this paper, we will study how to effectively apply the technique of directing through the composition of this shot to 360 degrees virtual reality. Currently, there are no killer contents that are still dominant in the world, including inside and outside the country. In this situation, the potential of virtual reality is recognized and various images are produced. So the way of production follows the traditional image production method, and the shot composition is the same. However, in the 360 degree virtual reality, the use of the long take or blocking technique of the conventional third person view point is used as the main production configuration, and the limit of the short configuration is felt. In addition, while the viewer can interactively view the 360-degree screen using the HMD tracking, the configuration of the shot and the connection of the shot are absolutely dependent on the director like the existing cinematography. In this study, I tried to study whether the viewer can freely change the cinematography such as the composition of the shot at a user's desired time using the feature of interaction of the VR image. To do this, 3D animation was created using a game tool called Unreal Engine to construct an interactive image. Using visual scripting of Unreal Engine called blueprint, we create a device that distinguishes the true and false condition of a condition with a trigger node, which makes a variety of shorts. Through this, various direction techniques are developed and related research is expected, and it is expected to help the development of 360 degree VR image.