• Journal of Digital Convergence
• /
• v.12 no.12
• /
• pp.521-528
• /
• 2014

In this paper, we propose the method of converting 2D image to 3D image using linear curves in HTML5. We use only one image without any other information about depth map for creating 3D images. So we filter the original image to extract RGB colors for left and right eyes. After selecting the ready-made control point of linear curves to set up depth values, users can set up the depth values and modify them. Based on the depth values that the end users select, we reflect them. Anaglyph 3D is automatically made with the whole and partial depth information. As all of this work has been designed and implemented in Web environment using HTML5, it is very easy and convenient and end users can create any 3D image that they want to make.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.5
• /
• pp.211-220
• /
• 2022

Recently, research on smart factories triggered by the 4th industrial revolution is being actively conducted. Accordingly, the manufacturing industry is conducting various studies to improve productivity and quality based on deep learning technology with robust performance. This paper is a study on the method of detecting tire surface defects in the visual inspection stage of the tire manufacturing process, and introduces a tire surface defect detection method using a depth image acquired through a 3D camera. The tire surface depth image dealt with in this study has the problem of low contrast caused by the shallow depth of the tire surface and the difference in the reference depth value due to the data acquisition environment. And due to the nature of the manufacturing industry, algorithms with performance that can be processed in real time along with detection performance is required. Therefore, in this paper, we studied a method to normalize the depth image through relatively simple methods so that the tire surface defect detection algorithm does not consist of a complex algorithm pipeline. and conducted a comparative experiment between the general normalization method and the normalization method suggested in this paper using YOLO V3, which could satisfy both detection performance and speed. As a result of the experiment, it is confirmed that the normalization method proposed in this paper improved performance by about 7% based on mAP 0.5, and the method proposed in this paper is effective.

• Broadcasting and Media Magazine
• /
• v.27 no.2
• /
• pp.43-50
• /
• 2022

한 장의 이미지로부터 장면의 깊이 정보를 추정하는 기술은 자율 주행, 실내외 로봇 기반 서비스 등 다양한 응용 분야에서 널리 적용되고 있다. 심층 학습을 이용한 알고리즘이 활발히 연구되면서 이러한 단안 깊이 추정 기술의 산업 분야 적용 범위는 확대되고 있는 추세이다. 그러나, 깊이 경계 정보를 정밀하게 예측하는데 여전히 많은 어려움이 있으며, 다양한 실제 환경에서 획득한 3차원 깊이 정보 구축 또한 많은 비용이 소모되는 문제점이 있다. 본 고에서는 이러한 문제를 해결하기 위해 최근 활발히 연구되고 있는 심층신경망 기반 단안 깊이 추정 연구의 최신 동향을 소개하고자 한다. 지도 학습 기반 방법부터 최근 활발히 연구되고 있는 비지도 학습 방법까지 상세히 살펴본다. 이와 더불어 대표 방법에 대한 성능 평가 결과도 간략히 제시하고자 한다.

• Journal of Broadcast Engineering
• /
• v.26 no.3
• /
• pp.283-295
• /
• 2021

This paper proposes a method for generating dense depth map using information of color images and depth map generated based on lidar based on self-organizing map. The proposed depth map upsampling method consists of an initial depth prediction step for an area that has not been acquired from LiDAR and an initial depth filtering step. In the initial depth prediction step, stereo matching is performed on two color images to predict an initial depth value. In the depth map filtering step, in order to reduce the error of the predicted initial depth value, a self-organizing map technique is performed on the predicted depth pixel by using the measured depth pixel around the predicted depth pixel. In the process of self-organization map, a weight is determined according to a difference between a distance between a predicted depth pixel and an measured depth pixel and a color value corresponding to each pixel. In this paper, we compared the proposed method with the bilateral filter and k-nearest neighbor widely used as a depth map upsampling method for performance comparison. Compared to the bilateral filter and the k-nearest neighbor, the proposed method reduced by about 6.4% and 8.6% in terms of MAE, and about 10.8% and 14.3% in terms of RMSE.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.826-829
• /
• 2024

3D 사람 자세 추정 기술은 다양한 응용 분야에서의 높은 활용성으로 인해 대량의 학습 데이터가 수집되어 딥러닝 모델 연구가 진행되어 온 반면, 동물 자세 추정의 경우 3D 동물 데이터의 부족으로 인해 관련 연구는 극히 미진하다. 본 연구는 동물 자세 추정을 위한 예비연구로서, 3D 학습 데이터가 없는 상황에서 단일 이미지로부터 3D 사람 자세를 추정하는 딥러닝 기법을 제안한다. 이를 위하여 사전 훈련된 다중 시점 학습모델을 사용하여 2D 자세 데이터로부터 가상의 다중 시점 데이터를 생성하여 훈련하는 연역적 학습 기반 교사-학생 모델을 구성하였다. 또한, 키포인트 깊이 정보 대신 2D 이미지로부터 레이블링 된 순서 깊이 정보에 기반한 손실함수를 적용하였다. 제안된 모델이 동물데이터에서 적용 가능한지 평가하기 위해 실험은 사람 데이터를 사용하여 이루어졌다. 실험 결과는 제안된 방법이 기존 단안 이미지 기반 모델보다 3D 자세 추정의 성능을 개선함을 보여준다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.319-321
• /
• 2020

스테레오 이미지는 카메라 두 대를 사용하여 찍은 미묘하게 다른 두 사진을 의미한다. 이 차이는 우리가 두 눈의 차이로 대상과의 거리를 계산하는 것처럼 이미지의 깊이를 예측할 수 있기에 요즘 주목받고 있는 자율주행 자동차나 로봇 비전에 이 기술이 사용된다. 심지어는 스마트폰 또한, 듀얼 카메라라는 이름으로 우리 일상생활에 쉽게 접할 수 있다. 이러한 상황에서 본 논문에서는 두 대의 카메라에서 찍힌 스테레오 이미지에서 손상된 한쪽의 이미지를 복원하는 기술문제에 대해 다루고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.10a
• /
• pp.1673-1674
• /
• 2015

본 논문에서는 깊이 카메라를 이용하여 사용자 수에 상관없이 사용자의 머리를 추적하는 방법에 대해 제안한다. 제안된 방법은 색상 정보를 제외한 깊이 정보만을 이용하여 머리를 추적하고, 각각의 사용자에 따라 깊이 이미지 형태가 다르게 나오는 머리를 실험적 데이터를 통하여 추적한다. 또한 제안된 방법은 카메라의 종류에 상관없이 머리를 추적할 수 있다는 장점이 있다. 본 논문에서는 Microsoft사의 Kinect for Window와 SoftKinetic사의 DS311을 실험을 진행하였다.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.26 no.4
• /
• pp.57-64
• /
• 2022

Chloride, which is one of the main deterioration factors in reinforced concrete structures, can degrade the performance of the structure due to chloride-induced corrosion of steel. Chloride content at steel depth or the rate of chloride penetration is necessary to determine deterioration of reinforced concrete or to calculate initiation time of steel corrosion caused by chloride attack. Chlorides in concrete are generally identified with typical two methods including chloride profiling using potentiometric titration method and discoloration method using AgNO₃ solution. The former is advantageous to estimate chloride penetration rate (diffusion coefficient in general) with measured chloride contents directly, but it is laborious. In the case of latter, while the result is obtained easily with the range of discoloration, the error may occur depending on workmanship when the depth of chloride ingress is measured. This study shows that chloride penetrated depth is evaluated with the results obtained from discoloration method through image analysis, thereby the error is minimized by workmanship. In addition, the effect of micro-crack in concrete is studied on chloride penetration. In conclusion, the depth of chloride penetration was quantified with image analysis and as it was confirmed that chlorides can rapidly penetrate through micro-cracks, caution is especially required for cracks in concrete structure.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.2
• /
• pp.47-54
• /
• 2011

This paper describes the method to reduce the time-cost for depth estimation of an image by learning, on the basis of the Genetic Algorithm, the image's features. The depth information is estimated from the relationship among features such as the energy value of an image and the gradient of the texture etc. The estimation-time increases due to the large dimension of an image's features used in the estimating process. And the use of the features without consideration of their importance can have an adverse effect on the performance. So, it is necessary to reduce the dimension of an image's features based on the significance of each feature. Evaluation of the method proposed in this paper using benchmark data provided by Stanford University found that the time-cost for feature extraction and depth estimation improved by 60% and the accuracy was increased by 0.4% on average and up to 2.5%.

• Journal of Advanced Navigation Technology
• /
• v.27 no.2
• /
• pp.182-189
• /
• 2023

Depth estimation is a key technology in 3D map generation for autonomous driving of vehicles, robots, and drones. The existing sensor-based method has high accuracy but is expensive and has low resolution, while the camera-based method is more affordable with higher resolution. In this study, we propose self-attention-based unsupervised monocular depth estimation for UAV camera system. Self-Attention operation is applied to the network to improve the global feature extraction performance. In addition, we reduce the weight size of the self-attention operation for a low computational amount. The estimated depth and camera pose are transformed into point cloud. The point cloud is mapped into 3D map using the occupancy grid of Octree structure. The proposed network is evaluated using synthesized images and depth sequences from the Mid-Air dataset. Our network demonstrates a 7.69% reduction in error compared to prior studies.