• Journal of the Korea Computer Graphics Society
• /
• v.26 no.3
• /
• pp.61-68
• /
• 2020

In this paper, we propose a deep learning framework for predicting a depth map of a 360° panorama image. Previous works use synthetic 360° panorama datasets to train networks due to the lack of realistic datasets. However, the synthetic nature of the datasets induces features extracted by the networks to differ from those of real 360° panorama images, which inevitably leads previous methods to fail in depth prediction of real 360° panorama images. To address this gap, we use domain adaptation to learn features shared by real and synthetic panorama images. Experimental results show that our approach can greatly improve the accuracy of depth estimation on real panorama images while achieving the state-of-the-art performance on synthetic images.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.3
• /
• pp.389-395
• /
• 2022

Because of improvement of deep learning techniques, deep learning using computer vision such as classification, detection and segmentation has also been used widely at many fields. Expecially, automatic driving is one of the major fields that applies computer vision systems. Also there are a lot of works and researches to combine multiple tasks in a single network. In this study, we propose the network that predicts the individual depth of pedestrians and vehicles. Proposed model is constructed based on YOLOv3 for object detection and Monodepth for depth estimation, and it process object detection and depth estimation consequently using encoder and decoder based on hard parameter sharing. We also used attention module to improve the accuracy of both object detection and depth estimation. Depth is predicted with monocular image, and is trained using self-supervised training method.

• The Journal of the Korea Contents Association
• /
• v.18 no.11
• /
• pp.447-454
• /
• 2018

This paper proposes a depth image generation algorithm of stereo images using a deep learning model composed of a CNN (convolutional neural network). The proposed algorithm consists of a feature extraction unit which extracts the main features of each parallax image and a depth learning unit which learns the parallax information using extracted features. First, the feature extraction unit extracts a feature map for each parallax image through the Xception module and the ASPP(Atrous spatial pyramid pooling) module, which are composed of 2D CNN layers. Then, the feature map for each parallax is accumulated in 3D form according to the time difference and the depth image is estimated after passing through the depth learning unit for learning the depth estimation weight through 3D CNN. The proposed algorithm estimates the depth of object region more accurately than other algorithms.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.105-106
• /
• 2018

본 논문에서는 동적 프로젝션 맵핑을 구현하기 위하여 3차원 공간의 깊이 정보와 대상 객체의 색상영상에서의 특징점을 추출하여 3차원 공간상에서 움직이는 2차원 평면 객체의 자세를 안정적으로 추정하는 기법을 제안한다. 제안 기법은 타겟 이미지를 출력하여 타겟 이미지 보다 큰 평면 패널에 부착하고, 이 평면 패널을 3차원 공간상에서 움직이는 환경에서 타겟 이미지의 자세를 안정적으로 추정하기 위하여 고안되었다. 제안 기법에서는 우선 패널이 움직일 수 있는 깊이 영역을 지정하여 해당 깊이 영역에 존재하는 2차원 패널을 추출하고, 패널의 사각영역을 추출한다. 또한, 색상 영상에 SURF 알고리즘을 적용하여 2차원 평면상에 부착된 타겟 이미지의 영역을 색상 특징을 기반으로 함께 추출하여 패널의 사각 영역과 타겟 이미지의 상대적인 위치 정보를 추출한다. 셋업 단계에서 추출된 타겟 이미지의 상대적인 위치 정보를 이용하여, 조명의 변화에 의하여 순간적으로 타겟 이미지의 특징점 추적에 실패한 경우, 패널의 사각 영역에 의해 계산된 타겟 이미지의 상대적 위치 정보를 계산하여 자세 추정에 사용함으로써 움직이는 타겟 이미지의 3차원 자세를 안정적으로 추정할 수 있도록 하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.112-115
• /
• 2021

최근 가상현실 및 증강 현실에 대한 관심도가 높아지면서, 깊이 추정, 객체 인식, 영상 분할 등의 다양한 컴퓨터 비전 알고리즘을 360° 영상에 적용하는 연구가 활발히 진행되고 있다. 이 중, 다수의 RGB 카메라를 활용하여 3 차원 정보를 추출하는 깊이 추정 기술은 보다 나은 몰입감을 제공하기 위한 핵심 기술이다. 그러나 깊이 추정 알고리즘의 객관적 성능 평가를 위한 정제된 360° 영상 데이터셋은 극히 부족하며, 이로 인하여 관련 분야 연구에 한계가 있다. 따라서 본 논문에서는 객관적인 알고리즘 성능 평가가 가능하며, 정제된 360° 동영상 데이터셋을 제안하고, 추후 다양한 360° 영상 응용 알고리즘 개발에 활용하고자 한다.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.1
• /
• pp.120-127
• /
• 2019

Passive source depth discrimination has been studied for decades since the source depth can be used for discriminating whether the target is near the surface or submerged. In this thesis, an algorithm for source depth discrimination is proposed based on CIR (Channel Impulse Response) from target-radiated noise (or signal). In order to extract CIR without a known source signal, Ray-based blind deconvolution is used. Subsequently, intersections of CIR pattern, which is characterized by ray arrival time difference, is utilized for discriminating source depth. The proposed algorithm is demonstrated through numerical simulation in ocean waveguide, and verified via the experimental data.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2013.06a
• /
• pp.289-292
• /
• 2013

최근 많은 영상처리 연구자들 사이에서 마이크로소프트사의 실시간 깊이센서 '키넥트'가 상당한 관심을 받고 있다. '키넥트'는 실시간으로 깊이정보를 제공함과 동시에 별도의 센서를 부착하지 않고도 컴퓨터와의 인터렉션할 수 있는 가능성을 제공한다. 하지만 '키넥트'의 깊이영상은 홀 영역, 부정확한 경계, 낮은 해상도등의 많은 문제점을 지니고 있다. 이러한 부정확한 깊이 정보는 3차원 렌더링, 가상시점 영상 합성, 모션 인식 등에서 성능 저하를 야기한다. 따라서 본 논문에서는 깊이 정보 품질 향상기법에 관하여 깊이영상 신뢰도를 이용한 도메인 변환기반 해상도 상향 알고리듬을 제안한다. 정확하고 빠르게 홀 영역정보를 추정하기 위해 도메인 변환 기반의 경계 보존 필터링이 사용된다. 또한 다양한 깊이 영상의 노이즈를 효율적으로 제거하기 깊이 영상의 신뢰도를 이용한다. 실험결과를 통하여 제안하는 방법이 효율적으로 홀 영역을 채우고, 부정확한 경계를 제거하여 깊이 영상의 품질을 향상시키는 것을 확인할 수 있다.

• Journal of Internet Computing and Services
• /
• v.19 no.5
• /
• pp.43-54
• /
• 2018

Images taken in haze weather are characteristic of low contrast and poor visibility. The process of reconstructing clear-weather image from a hazy image is called dehazing. The main challenge of image dehazing is to estimate the transmission map or depth map for an input hazy image. In this paper, we propose a single image dehazing method by utilizing the Generative Adversarial Network(GAN) for accurate depth map estimation. The proposed GAN model is trained to learn a nonlinear mapping between the input hazy image and corresponding depth map. With the trained model, first the depth map of the input hazy image is estimated and used to compute the transmission map. Then a guided filter is utilized to preserve the important edge information of the hazy image, thus obtaining a refined transmission map. Finally, the haze-free image is recovered via atmospheric scattering model. Although the proposed GAN model is trained on synthetic indoor images, it can be applied to real hazy images. The experimental results demonstrate that the proposed method achieves superior dehazing results against the state-of-the-art algorithms on both the real hazy images and the synthetic hazy images, in terms of quantitative performance and visual performance.

• Journal of Broadcast Engineering
• /
• v.17 no.4
• /
• pp.611-624
• /
• 2012

'Gesture' is the most intuitive means of communication except the voice. Therefore, there are many researches for method that controls computer using gesture input to replace the keyboard or mouse. In these researches, the method of user detection and main body parts estimation is one of the very important process. in this paper, we propose user objects detection and main body parts estimation method on inaccurate depth information for pose estimation. we present user detection method using 2D and 3D depth information, so this method robust to changes in lighting and noise and 2D signal processing 1D signals, so mainly suitable for real-time and using the previous object information, so more accurate and robust. Also, we present main body parts estimation method using 2D contour information, 3D depth information, and tracking. The result of an experiment, proposed user detection method is more robust than only using 2D information method and exactly detect object on inaccurate depth information. Also, proposed main body parts estimation method overcome the disadvantage that can't detect main body parts in occlusion area only using 2D contour information and sensitive to changes in illumination or environment using color information.

• International Journal of Highway Engineering
• /
• v.2 no.1
• /
• pp.135-145
• /
• 2000

A new procedure needs to be developed to predict the dynamic layer properties under moving truck loads. In this study, a computer code to evaluate layer moduli of asphalt concrete pavement from measured interior deflections at various depths were developed and verified from numerical model tests. Interior deflections of the pavement are measured from Multi-Depth Deflectometer(MDD). It was found that errors between the given and backcalculated moduli in numerical analysis were less than 0.32% for several numerical models tested. When impact loads were used, a technique to determine the depth to virtual rigid base was proposed through the analysis of compressive wave velocity and impulse loading durations. It was found that errors between the given and backcalculated moduli in numerical analysis were less than 0.114% when virtual rigid base was considered in numerical analysis. The pavement behavior must be evaluated under various vehicle speeds when determining the dynamic interaction between the loading vehicle and pavement system. To evaluate the dynamic behavior on asphalt concrete pavement under various vehicle speeds, truck moving tests were carried out. From the test results with respect to vehicle speed, it was found that the vehicle speed had significant effect on actual response of the pavement system. The lower vehicle speed generates the higher interior deflections, and the lower dynamic modulus.