http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Real-Time 2D-to-3D Conversion for 3DTV using Time-Coherent Depth-Map Generation Method
남승우,김혜선,반윤지,진성일 한국콘텐츠학회 2014 International Journal of Contents Vol.10 No.3
Depth-image-based rendering is generally used in real-time 2D-to-3D conversion for 3DTV. However, inaccurate depth maps cause flickering issues between image frames in a video sequence, resulting in eye fatigue while viewing 3DTV. To resolve this flickering issue, we propose a new 2D-to-3D conversion scheme based on fast and robust depth-map generation from a 2D video sequence. The proposed depth-map generation algorithm divides an input video sequence into several cuts using a color histogram. The initial depth of each cut is assigned based on a hypothesized depth-gradient model. The initial depth map of the current frame is refined using color and motion information. Thereafter, the depth map of the next frame is updated using the difference image to reduce depth flickering. The experimental results confirm that the proposed scheme performs real-time 2D-to-3D conversions effectively and reduces human eye fatigue.
Hole-Filling Methods Using Depth and Color Information for Generating Multiview Images
남승우,장경호,반윤지,김혜선,진성일 한국전자통신연구원 2016 ETRI Journal Vol.38 No.5
This paper presents new hole-filling methods for generating multiview images by using depth image based rendering (DIBR). Holes appear in a depth image captured from 3D sensors and in the multiview images rendered by DIBR. The holes are often found around the background regions of the images because the background is prone to occlusions by the foreground objects. Background-oriented priority and gradient-oriented priority are also introduced to find the order of hole-filling after the DIBR process. In addition, to obtain a sample to fill the hole region, we propose the fusing of depth and color information to obtain a weighted sum of two patches for the depth (or rendered depth) images and a new distance measure to find the best-matched patch for the rendered color images. The conventional method produces jagged edges and a blurry phenomenon in the final results, whereas the proposed method can minimize them, which is quite important for high fidelity in stereo imaging. The experimental results show that, by reducing these errors, the proposed methods can significantly improve the hole-filling quality in the multiview images generated.
효과적인 Fur 렌더링을 위한 적응적 시스템 : 혼합 렌더링을 이용한 빠른 Fur 렌더링 방법
김혜선(Hyesun Kim),반윤지(Yunji Ban),이충환(Chunghwan Lee),남승우(Seungwoo Nam),최진성(Jinsung Choi),오준규(Junkyu Oh) 한국HCI학회 2009 한국HCI학회 학술대회 Vol.2009 No.2
Fur rendering is difficult in that there are huge numbers of objects and it takes so much time. The previous method considers fur as cylinder, transforms it into 2D ribbon, triangulates and commits rendering. But this method has problem like under sampling and takes rendering time so long. To resolve these shortcuts we proposed new algorithm. We divide fur into thick and thin fur and we applied adaptive rendering methods for each type of fur. Also we can perform an effective rendering according to the proposed rendering framework. 털 렌더링은 사실적인 렌더링을 위해 많은 수의 털 데이터를 처리해야 하는 어려움이 있다. 대량의 털 데이터를 렌더링함에 있어서 가장 어려운 점은 렌더링 시간이 많이 걸린다는 점이다. 기존의 털 렌더링 방법은 털을 원통형의 실린더로 간주하고 2D 형태의 리본으로 변환하고 삼각화하여 렌더링하는 방법이다. 하지만 이 방법은 언더 샘플링(under sampling)문제가 있고 렌더링 시간이 오래 걸린다는 단점이 있다. 이런 단점을 개선하기 위해서 이 논문에서는 새로운 알고리즘을 제안하였다. 털을 굵기에 따라 나누고 굵은 털과 가는 털에 각각의 렌더링 방법을 사용함으로써 렌더링 속도를 개선하였다. 또한 전체 렌더링 프레임워크에 대한 제안을 함으로써 보다 효과적인 렌더링을 수행할 수 있다.
박지원(Jiwon Park),이세형(Seihyoung Lee),반윤지(Yun-Ji Ban),민기현(Gihyeon Min),김정은(Jeongeun Kim) 한국정보기술학회 2024 한국정보기술학회논문지 Vol.22 No.2
A classification model using deep learning was developed for automatic diagnosis of diabetic foot ulcer disease. In this study, we implemented an automated medical image disease diagnosis model that can classify diabetic foot ulcers into normal and diseased, and classify them into infection and ischemia, complex ulcers, and other ulcers by using a publicly available image dataset of diabetic foot ulcers(DFU challenge 2021). For this purpose, six deep neural network models were used to train, and the accuracy, recall, precision, and f1-score performance of the models were compared and evaluated for each model to present a model with performance suitable for diabetic foot ulcer classification. To improve the performance of model, data augmentation and imbalanced data processing were used, and among them, the EfficientNetB3 model achieved 89% accuracy and 89% recall for diabetic foot ulcer data.