http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
이주성(Jusung Lee),임정훈(Jeonghun Lim),백승엽(Stephen S. Beak),이건우(Kunwoo Lee) 대한기계학회 2014 대한기계학회 춘추학술대회 Vol.2014 No.11
In Dental CAD, It is Important to extract exactly margin Line from teeth model. In particular, the teeth have a variety of structures of the scan data for each patient. There is a need for a method readily available to the margin line. Therefore, we developed the active contour model to generate on the mesh. The margin line extraction method requires a minimum of user interaction.
ELK 스택을 활용한 로그 기반 콘텐츠 동시 시청 제어 실증
이주성(Lee Jusung),신승호(Shin Seungho),이상범(Lee Sangbeom) 한국통신학회 2021 한국통신학회 학술대회논문집 Vol.2021 No.6
본 논문에서는 오픈소스 솔루션인 ELK 를 사용하여 Log 기반 스트리밍 컨텐츠의 동시 시청 제어를 수행한다. 시청 Log 적재는 Kafka 를 사용하고, 시청 로그 분석에는 Elasticsearch 를 사용하여 동시 시청을 판별한다. 이를 실제 시스템에 적용하여 REST API latency 가 3ms 이하의 성능을 달성하였다.
이주성(Jusung Lee),백승엽(Seung-Yeob Baek),이건우(Kunwoo Lee) (사)한국CDE학회 2010 한국CDE학회 논문집 Vol.15 No.4
Medical image acquisition techniques such as CT and MRI have disadvantages in that the numerous time and efforts are needed. Furthermore, a great amount of radiation exposure is an inherent proberty of the CT imaging technique, a number of side-effects are expected from such method. To improve such conventional methods, a number of novel methods that can obtain 3D medical images from a few X-ray images, such as algebraic reconstruction technique (ART), have been developed. Such methods deform a generic model of the internal body part and fit them into the X-ray images to obtain the 3D model; the initial shape, therefore, affects the entire fitting process in a great deal. From this fact, we propose a novel method that can generate a 3D vertebraic generic model based on the statistical database of CT scans in this study. Moreover, we also discuss a method to generate patient-tailored generic model using the facts obtained from the statistical analysis. To do so, the mesh topologies of CTscanned 3D vertebra models are modified to be identical to each other, and the database is constructed based on them. Furthermore, from the results of a statistical analysis on the database, the tendency of shape distribution is characterized, and the modeling parameters are extracted. By using these modeling parameters for generating the patient-tailored generic model, the computational speed and accuracy of ART can greatly be improved. Furthermore, although this study only includes an application to the C1 (Atlas) vertebra, the entire framework of our method can be applied to other body parts generally. Therefore, it is expected that the proposed method can benefit the various medical imaging applications.
컴퓨터를 이용한 의료 진단용 3 차원 척추 제네릭 모델
이주성(Jusung Lee),백승엽(Seung-Yeob Baek),이건우(Kunwoo Lee) (사)한국CDE학회 2010 한국 CAD/CAM 학회 학술발표회 논문집 Vol.2010 No.1
CT 나 MRI 와 같은 기존 3 차원 의료 영상 획득방법은 많은 비용과 시간이 소모된다는 단점을 가지고 있다. 또한 CT 영상은 다른 방법들에 비하여 방사선 피폭량이 상대적으로 많으므로, 이로인한 부작용 또한 문제점으로 지적될 수 있다. 이러한 기존방법들의 단점을 개선하기 위하여 최근 ART(Algebraic Reconstruction Technique)와 같이 몇 장의 X-ray 영상 만으로 환자의 3 차원 영상을 획득할 수 있는 기술들이 개발된 바 있다. 이러한 기술들은 하나의 제네릭 모델을 X-ray 영상에 맞게 적절히 변형시킴으로써 구현될 수 있는데, 제네릭 모델의 초기 형상에 따라 계산 속도나 결과물의 형상 등이 달라질 수 있다. 이러한 사실에 착안하여 우리는 통계적 자료를 근거로 3 차원 척추(C1) 제네릭 모델을 생성하고, 이를 여러 환자 각각에 특화된 모델(patient-taylored model)로 변형시키는 방법을 제안한다. 우리는 이 논문을 통해 CT 촬영으로 얻은 여러 개의 3 차원 척추 모델 데이터들의 메쉬 토폴로지를 일치시키고, 이를 기반으로 통계 데이터베이스를 구축하는 방법에 대해 논의한다. 나아가 이러한 통계 데이터베이스에 대한 수학적인 분석을 통하여 척추 형상 분포의 경향성을 찾아내고, 이를 종합하여 손쉽게 환자에게 특화된 제네릭 모델을 생성하는 방법에 대해서도 논의할 것이다. 이러한 방법으로 환자에게 특화되어 생성된 제네릭 모델을 이용하게 되면 X-ray 영상으로부터 3 차원 모델을 생성함에 있어서 그 처리 속도 뿐만아니라 정확도 또한 증대시킬 수 있다. 또한 이러한 방법론은 본 논문에서 주로 초점을 맞추고 있는 C1(Atlas) 뿐 아니라 다른 부위의 뼈 형상들에 대해서도 동일하게 적용할 수 있기 때문에, 앞으로 의료 영상 분야에서 다양한 응용이 가능하리라 기대된다.
Initial closed curve for active contour model on mesh
JuSung Lee(이주성),JeongHun Lim(임정훈),Kunwoo Lee(이건우) (사)한국CDE학회 2015 한국 CAD/CAM 학회 학술발표회 논문집 Vol.2015 No.동계
The accurate margin line extraction is a crucial process in computer-aided orthodontics. Margin line is boundary line of treatment region. Though curvature field is traditionally used for detect feature line, different shapes and scanning noises make hard to detect margin line automatically. Thus current dental CAD systems mostly require the user interactive definition of points by the user. So, we also use interactive defined starting points by the user. In medical image processing, active contour model is used to extract feature line. However it doesn’t work when initial curve is improper. To solve this problem, this paper proposes a novel closed curve initialization algorithm that makes initial curve on mesh by using starting points. Some experiments show our initialization method is robust and fast.
이주성(Jusung Lee),이영현(Younghyun Lee),이준수(Joonsoo Lee) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.8
Scene text recognition is a challenging task because it contains a variety of background, noise, blur, fonts, and perspective views. Recently, deep learning-based algorithms, for example, RNNbased methods or Transformer-based methods, have shown outstanding results on the scene text recognition task. However, RNN-based methods suffer inherently from long-term dependency problems and Transformer-based methods have low accuracy in curved and irregular text. In this paper, we propose a novel deep neural network architecture which combines “Spatial Transformer Network” and “Transformer” network for scene text recognition. The proposed architecture shows consistently better word accuracy over widely used public word recognition datasets, compared to previous scene text recognition models.
이주성(Jusung Lee),백승엽(Seung-Yeob Baek),이건우(Kunwoo Lee) (사)한국CDE학회 2011 한국 CAD/CAM 학회 학술발표회 논문집 Vol.2011 No.1
Medical image acquisition techniques such as CT and MRI have disadvantages in that the numerous time and efforts are needed. Furthermore, a great amount of radiation exposure is an inherent proberty of the CT imaging technique, a number of side-effects are expected from such method. To improve such conventional methods, a number of novel methods that can obtain 3D medical images from a few X-ray images, such as algebraic reconstruction technique (ART), have been developed. Such methods deform a generic model of the internal body part and fit them into the X-ray images to obtain the 3D model; the initial shape, therefore, affects the entire fitting process in a great deal. From this fact, we propose a novel method that can generate a 3D vertebraic generic model based on the statistical database of CT scans in this study. Moreover, we also discuss a method to generate patient-tailored generic model using the facts obtained from the statistical analysis. To do so, the mesh topologies of CTscanned 3D vertebra models are modified to be identical to each other, and the database is constructed based on them. Furthermore, from the results of a statistical analysis on the database, the tendency of shape distribution is characterized, and the modeling parameters are extracted. By using these modeling parameters for generating the patient-tailored generic model, the computational speed and accuracy of ART can greatly be improved. Furthermore, although this study only includes an application to the C1 (Atlas) vertebra, the entire framework of our method can be applied to other body parts generally. Therefore, it is expected that the proposed method can benefit the various medical imaging applications.
이재명(Jaemyung Lee),이주성(Jusung Lee),이영현(Younghyun Lee),이준수(Jaemyung Lee) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.8
Instance segmentation-based approach using Mask R-CNN is currently one of the leading methods for the text localization task. Different from general object detection tasks, the aspect ratio of text instances is too high to apply Mask R-CNN as it is. To simply apply Mask R-CNN for text detection yields false positives due to overgeneralized receptive fields for bounding boxes. In this paper, we propose a modified Mask R-CNN architecture for text detection. We present a method to extract features containing word-level and character-level receptive fields simultaneously. Our approach shows consistent performance improvement on MLT 2017 and Incidental Scene Text. Moreover, our method surpasses most of prior state-of-the-art text localization methods appeared in recent computer vision conferences.