• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권11호
• /
• pp.5555-5567
• /
• 2018

State-of-art instance segmentation networks are successful at generating 2D segmentation mask for region proposals with highest classification score, yet 3D object segmentation task is limited to geocentric embedding or detector of Sliding Shapes. To this end, we propose an amodal 3D instance segmentation network called A3IS-CNN, which extends the detector of Deep Sliding Shapes to amodal 3D instance segmentation by adding a new branch of 3D ConvNet called A3IS-branch. The A3IS-branch which takes 3D amodal ROI as input and 3D semantic instances as output is a fully convolution network(FCN) sharing convolutional layers with existing 3d RPN which takes 3D scene as input and 3D amodal proposals as output. For two branches share computation with each other, our 3D instance segmentation network adds only a small overhead of 0.25 fps to Deep Sliding Shapes, trading off accurate detection and point-to-point segmentation of instances. Experiments show that our 3D instance segmentation network achieves at least 10% to 50% improvement over the state-of-art network in running time, and outperforms the state-of-art 3D detectors by at least 16.1 AP.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 1995년도 학술대회
• /
• pp.143-148
• /
• 1995

In this paper, we describe a fast 3D video segmentation method using mathematical morphology. The proposed 3D video segmentation algorithm is composed of intra-frame segmentation step and inter-frame segmentation step. In the intra-frame segmentation step, the first frame is segmented using the fast hierarchical segmentation method. Then, in the inter-frame segmentation step, the next frames are segmented using markers that are extracted from the difference of previous segmentation result and simplified present image. Experimental results show that the proposed method has more fast structure and is suitable for video segmentation.

• 한국인터넷방송통신학회논문지
• /
• 제11권3호
• /
• pp.129-135
• /
• 2011

본 논문에서는 2D 동영상을 3D 입체영상으로 변환하기 위해서 머신러닝에 의한 학습기반의 객체분할과 객체의 optical flow를 활용하는 방법을 제안한다. 성공적인 3D 변환을 가능하게 하는 객체분할을 위해서, 객체의 칼라 및 텍스쳐 정보는 학습을 통해 반영하고 움직임이 있는 영역 위주로 객체분할을 수행할 수 있도록 optical flow를 도입한 새로운 에너지함수를 설계하도록 한다. 분할된 객체들에 대해 optical flow 크기에 따른 깊이맵을 추출하여 입체영상에 필요한 좌우 영상을 합성하여 생성하도록 한다. 제안한 기법으로 인해 효과적인 객체분할과 깊이맵을 생성하여 2D 동영상에서 3D 입체동영상으로 변환됨을 실험결과들이 보여준다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 추계종합학술대회 논문집(3)
• /
• pp.177-180
• /
• 2000

CT images are sequential images that provide medical doctors helpful information for treatment and surgical operation. It is also widely used for the 3D reconstruction of human bone and organs. In the 3D reconstruction, the quality of the reconstructed 3D model heavily depends on the segmentation results. In this paper, we propose an algorithm suitable for the segmentation of teeth and the maxilofacial bone.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.55-58
• /
• 2009

Automatic segmentation of brain MRI data usually leaves some segmentation errors behind that are to be subsequently removed interactively, using computer graphics tools. This interactive removal is normally performed by operating on individual 2D slices. It is very tedious and still leaves some segmentation errors which are not visible on the slices. We have proposed to perform a novel 3D interactive correction of brain segmentation errors introduced by the fully automatic segmentation algorithms. We have developed the tool which is based on 3D semi-automatic propagation algorithm. The paper describes the implementation principles of the proposed tool and illustrates its application.

• 한국전자파학회논문지
• /
• 제11권4호
• /
• pp.582-591
• /
• 2000

본 논문에서는 휴대전화기에 의한 인체 영향을 FDTD (시간영역 유한차분법) 해석할 수 있도록 한국인 표준 에 알맞는 인체 두부의 FDTD 모텔 제작 방법을 소개하였다. 한국인 표준에 알맞은 사람의 두부를 MRI 촬영한 다음.2차원 MRI 영상 데이터를 이용하여 2차원 segmentation을 하였다. segmentation은 반자동법을 적용하였 으며 제작된 2차원 se밍nentation 데이터를 토대로 $1mm\times1mm\times1mm$크기의 3차원 고해상도 segmentation 데이터를 제작하였다. 3차원 고해상도 segmentation 데이터를 이용하여 휴대전화기의 사용 상황에 어올리도록 다양한 각도로 기울인 인체 두부의 FDTD 모델을 제작하였다.

• 한국산업융합학회 논문집
• /
• 제27권3호
• /
• pp.563-572
• /
• 2024

Image processing and computer vision technologies are becoming increasingly important in a variety of application fields that require techniques and tools for sophisticated image analysis. In particular, image segmentation is a technology that plays an important role in image analysis. In this study, in order to identify recent research trends on image segmentation techniques, we used the Web of Science(WoS) database to analyze the R&D topography based on the network structure of the author's keyword co-occurrence matrix. As a result, from 2015 to 2023, as a result of the analysis of the R&D map of research articles on image segmentation, R&D in this field is largely focused on four areas of research and development: (1) researches on collecting and preprocessing image data to build higher-performance image segmentation models, (2) the researches on image segmentation using statistics-based models or machine learning algorithms, (3) the researches on image segmentation for medical image analysis, and (4) deep learning-based image segmentation-related R&D. The scientometrics-based analysis performed in this study can not only map the trajectory of R&D related to image segmentation, but can also serve as a marker for future exploration in this dynamic field.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2001년도 하계종합학술대회 논문집(5)
• /
• pp.73-76
• /
• 2001

A segmentation is a prior processing for medical image analysis and 3D reconstruction. This Paper provides the method to segment 3D Visible Human color images. Firstly, the reference images that have a initial curve are segmented using Balloon and the results are propagated to the adjacent images. In the propagation processing, the result of the adjacent slice is modified by Edge-limited SRG Finally, the 3D Balloon improves the segmentation results of each 2D slice. the proposed method's performance was verified through the experiments to segment thigh muscles of Visible Human color images.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권3호
• /
• pp.1418-1433
• /
• 2019

In computer graphics, 3D mesh segmentation is a challenging research field. This paper presents a 3D mesh model segmentation algorithm that focuses on removing exterior salient parts from the original 3D mesh model based on prominent feature points and marching plane. To begin with, the proposed approach uses multi-dimensional scaling to extract prominent feature points that reside on the tips of each exterior salient part of a given mesh. Subsequently, a set of planes intersect the 3D mesh; one is the marching plane, which start marching from prominent feature points. Through the marching process, local cross sections between marching plane and 3D mesh are extracted, subsequently, its corresponding area are calculated to represent local volumes of the 3D mesh model. As the boundary region of an exterior salient part generally lies on the location at which the local volume suddenly changes greatly, we can simply cut this location with the marching plane to separate this part from the mesh. We evaluated our algorithm on the Princeton Segmentation Benchmark, and the evaluation results show that our algorithm works well for some categories.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 ITC-CSCC -1
• /
• pp.450-453
• /
• 2000

In this paper, an automated segmentation algorithm is proposed for MR brain images using T1-weighted, T2-weighted, and PD images complementarily. The proposed segmentation algorithm is composed of 3 steps. In the first step, cerebrum images are extracted by putting a cerebrum mask upon the three input images. In the second step, outstanding clusters that represent inner tissues of the cerebrum are chosen among 3-dimensional (3D) clusters. 3D clusters are determined by intersecting densely distributed parts of 2D histogram in the 3D space formed with three optimal scale images. Optimal scale image best describes the shape of densely distributed parts of pixels in 2D histogram. In the final step, cerebrum images are segmented using FCM algorithm with it’s initial centroid value as the outstanding cluster’s centroid value. The proposed segmentation algorithm complements the defect of FCM algorithm, being influenced upon initial centroid, by calculating cluster’s centroid accurately And also can get better segmentation results from the proposed segmentation algorithm with multi spectral analysis than the results of single spectral analysis.