• Journal of the Korean Society for Precision Engineering
• /
• v.10 no.4
• /
• pp.170-179
• /
• 1993

In this study, self-organized neural network is used to solve the vorrespondence problem of the axial stereo image. Edge points are extracted from a pair of stereo images and then the edge points of rear image are assined to the output nodes of neural network. In the matching process, the two input nodes of neural networks are supplied with the coordi- nates of the edge point selected randomly from the front image. This input data activate optimal output node and its neighbor nodes whose coordinates are thought to be correspondence point for the present input data, and then their weights are allowed to updated. After several iterations of updating, the weights whose coordinates represent rear edge point are converged to the coordinates of the correspondence points in the front image. Because of the feature map properties of self-organized neural network, noise-free and smoothed depth data can be achieved.

• Journal of Advanced Navigation Technology
• /
• v.13 no.1
• /
• pp.142-150
• /
• 2009

It is very important to generate a simplified model from a complex 3D character in computer game. We propose a new method of extracting feature lines from a 3D game character. Given an unstructured 3D character model containing texture information, we use model feature map (MFM), which is a 2D map that abstracts the variation of texture and curvature in the 3D character model. The MFM is created from both a texture map and a curvature map, which are produced separately by edge-detection to locate line features. The MFM can be edited interactively using standard image-processing tools. We demonstrate the technique on several data sets, including, but not limited to facial character.

• Proceedings of the Korea Contents Association Conference
• /
• 2009.05a
• /
• pp.152-156
• /
• 2009

The goal of this study is to represent heightfield-map based cartoon liquid simulation using GPU. To do this, we examine the optical and flow features of liquid. So far, NPR represent study for gas and liquid advanced in computer graphics. In this study the flow of fluid represent using a trivial algorithm using a non-physical method, and optical of liquid represent ratio of value of reflection and refraction using fresnel equation. Cartoon represent of liquid different from the existing method, that the value of reflection and refraction on the environment map. Edge detection represent jump edge of liquid. As a result, the liquid simulation could represent a cartoon.

• Proceedings of the KSRS Conference
• /
• v.2
• /
• pp.643-646
• /
• 2006

This paper addresses the approach to extract linear features from satellite imagery using an efficient segmentation method. The extraction of linear features from satellite images has been the main concern of many scientists. There is a need to develop a more capable and cost effective method for the Iranian map revision tasks. The conventional approaches for producing, maintaining, and updating GIS map are time consuming and costly process. Hence, this research is intended to investigate how to obtain linear features from SPOT satellite imagery. This was accomplished using a discontinuity-based segmentation technique that encompasses four stages: low level bottom-up, middle level bottom-up, edge thinning and accuracy assessment. The first step is geometric correction and noise removal using suitable operator. The second step includes choosing the appropriate edge detection method, finding its proper threshold and designing the built-up image. The next step is implementing edge thinning method using mathematical morphology technique. Lastly, the geometric accuracy assessment task for feature extraction as well as an assessment for the built-up result has been carried out. Overall, this approach has been applied successfully for linear feature extraction from SPOT image.

• Spatial Information Research
• /
• v.11 no.4
• /
• pp.421-437
• /
• 2003

In order to calculate the position of parcel boundary points, the area of the Parcel, and the length of the parcel boundary lines using a digitalized cadastral map or maps, the distortion of the map has to be corrected. The correction methods, such as the method of 2D affine transformation using 4 comer points of the edge lines, the method of tessellation into 9 subregions, the method of using the original surveying sheets, and the method of straightening the 4 edge lines, have been developed. In this Paper, the four methods were programed and applied to some sample cadastral maps and the correction accuracies were obtained and analysed. No method could prefectly correct the distortions because the distortions were irregular throughout the maps. However, it is found that tile method of straightening the 4 edge lines is the one which can minimize the distortions when the method is applied after applying the 2D projective transformation on the maps using the 4 comer points of the edges.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.5
• /
• pp.2529-2543
• /
• 2019

Scanning electron microscopy (SEM) image can link with the microscopic world through reflecting interaction between electrons and materials. The SEM images are easily subject to blurring distortions during the imaging process. Inspired by the fact that dark channel prior captures the changes to blurred SEM images caused by the blur process, we propose a method to evaluate the SEM images sharpness based on the dark channel prior. A SEM image database is first established with mean opinion score collected as ground truth. For the quality assessment of the SEM image, the dark channel map is generated. Since blurring is typically characterized by the spread of edge, edge of dark channel map is extracted. Then noise is removed by an edge-preserving filter. Finally, the maximum gradient and the average gradient of image are combined to generate the final sharpness score. The experimental results on the SEM blurred image database show that the proposed algorithm outperforms both the existing state-of-the-art image sharpness metrics and the general-purpose no-reference quality metrics.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.17 no.6
• /
• pp.744-748
• /
• 2007

In this paper, a new localization method of indoor mobile robot using multi IR(infrared) range sensors is proposed. Each IR range sensor detects the edge of obstacles and wall using the acquired range data. The environment map is built by the merging process of the detected edge data of each sensor. The performance of proposed system is verified by the comparison of the real environment and the detected map in experiments.

• Korean Journal of Computational Design and Engineering
• /
• v.3 no.3
• /
• pp.161-167
• /
• 1998

Cutting forces in ball-end milling of slanted surfaces are calculated. The cutting area is determined from the Z-map of the surface geometry and current cutter location. The obtained cutting area is projected onto the cutter plane normal to the Z-axis and compared with cutting edge element location. Cutting force is calculated by integration of elemental cutting forces of engaged cutting edge elements. Experiments with various slanted angles were performed to verify the proposed cutting force estimation model. It is shown that the proposed method predicts cutting force effectively for any geometry including sculptured surfaces with cusp marks and surfaces with pockets and holes.

• Journal of Korea Game Society
• /
• v.20 no.3
• /
• pp.65-74
• /
• 2020

In the shadow map method, if the shadow map is magnified, the shadow has a jagged silhouette. Herein, we propose a soft shadow method that filters reshaped silhouettes analytically. First, the shadow silhouette is reshaped through sub-texel edge detection, which is based on linear or quadratic curve models. Second, an integral shadow filtering algorithm is used to accurately obtain the average shadow intensity from a definite integral estimation. The implementation demonstrates that our solution can effectively eliminate jagged aliasing and efficiently generate soft shadows.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.695-696
• /
• 2020

CNN(CNN: Convolutional Neural Network)은 컴퓨터 비전의 많은 분야에서 뛰어난 성능을 보이고 있으며, 단일 영상으로부터 깊이(depth) 추정에서도 기존 기법보다 향상된 성능을 보이고 있다. 그러나, 단일 영상으로부터 신경망이 얻을 수 있는 정보는 제한적이기 때문에 스테레오 카메라로부터 얻은 좌/우 영상으로부터의 깊이 추정보다 성능 향상에 한계가 있다. 따라서 본 논문에서는 에지 맵(edge map)을 이용한 CNN 기반의 단일 영상에서의 깊이 추정의 개선 기법을 제안한다. 제안 방법은 먼저 단일 영상에 대한 전처리를 통해서 에지 맵과 양방향 필터링된(bilateral filtered) 영상을 생성하고, 이를 CNN 입력으로 하여 기존 단일 영상 깊이 추정 기법 대비 개선된 성능을 보임을 확인하였다.