• 한국콘텐츠학회논문지
• /
• 제6권9호
• /
• pp.19-27
• /
• 2006

Mean shift 방법은 중심 모드를 찾기 위한 비모수적 통계 방법으로 컬러 영상을 분할하는데 효율적이다. 그러나 입력되는 윈도우 크기에 따라 분할된 결과가 달라지며 윈도우의 크기 값이 작을 경우 많은 영역으로 분할되는 단점이 있다. 본 논문은 이러한 단점을 개선하여 mean shift 알고리즘에 의한 분할 영상이 과도하게 분할되었을 경우 영역 병합 방법을 이용하여 유사 영역을 병합하는 방법을 제안한다. 제안 방법은 과분할된 영상을 HSI 컬러 공간으로 변환하여 색상 정보를 이용하여 유사 영역으로 병합하며 이때 경계 영역을 보존하기 위해 영역 병합 제한자를 이용하여 병합 유무를 결정한다. 그 후 RGB 컬러 공간을 이용하여 HSI 컬러 공간에서 병합되지 않은 영역들을 병합하였다. 실험 결과는 다양한 영상에 대해 주요 영역들의 분할 결과에서 우수한 성능을 보여준다.

• 한국수소및신에너지학회논문집
• /
• 제21권1호
• /
• pp.1-11
• /
• 2010

Current distribution measurement technique based on a segmented bipolar plate (BP) has been widely adopted to visualize the distribution of current density in a polymer electrolyte membrane. However, a concern is raised how closely the current density of a segmented BP can approach that of a corresponding non-segmented membrane. Therefore, in this paper, the accuracy of the measurement technique is numerically evaluated by applying a three-dimensional, two-phase fuel cell model to a $100\;cm^2$ area fuel cell geometry in which segmented BPs and non-segmented membrane are combined together. The simulation results reveal that the errors between the current densities of the segmented BPs and non-segmented membrane indeed exist, predicting the maximum relative error of 33% near the U-turn regions of the flow-field. The numerical study further illustrates that the erroneous result originates from the BPs segmented non-symmetrically based on the flow channels that allows some currents bypassing flow channels to flow into its neighboring segment. Finally, this paper suggests the optimal way for bipolar plate segmentation that can minimize the deviation of current measured in a segmented BP from that of a corresponding membrane region.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제34권10호
• /
• pp.929-936
• /
• 2007

본 논문에서는 교통 모니터링 시스템에 사용할 수 있는 국부 영역에서 차량 검지와 추적을 수행하는 새로운 기법을 제안하다. 차량 검지와 추적은 각 차선에 미리 설정된 영역에서만 이루어진다. 각 차선에 설정된 국부 영역을 에지 특성과 프레임 차이를 이용하여 여러 개의 분할 영역으로 나누고 분할영역의 통계적 특성과 기하학적 특성에 의해 차량, 도로, 그림자와 전조등 영역으로 분류하여 차량을 검출한다. 검출된 차량은 에지 영상의 정합에 의해 국부 영역내에서 추적하여 차량 속도, 길이, 차간 거리와 도로 점유율과 같은 교통 정보를 산출할 수 있다. 배경 영상을 사용하지 않으므로 다양한 조건에서 사용이 가능하고 다양한 기상, 시간대와 장소에서 90.16%의 높은 차량 검출의 정확도를 나타냈다. 동작 환경에서 카메라의 각도, 방향과 조리개 설정이 조정되면 아주 높은 정확도의 교통 모니터링 시스템의 핵심기술로 사용될 수 있을 것으로 기대된다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1998년도 추계종합학술대회 논문집
• /
• pp.915-918
• /
• 1998

Region based coding consistsof image segmentation contour and texture coding. Contour coding techniques can be classified into contour or shape-oriented approaches. In this paper, geodesic skeleton based on shape-oriented approach is used for contour coding. Efficient application of geodesic skeleton for contour coding based on the characteristics of regions in segmented image will be discussed.

• 대한의용생체공학회:의공학회지
• /
• 제10권3호
• /
• pp.317-322
• /
• 1989

In this paper, brain CT images are automatically segmented to reconstruct the 3-D scene from consecutive CT sections. Contextual segmentation technique was applied to overcome the partial volume artifact and statistical fluctuation phenomenon of soft tissue images. Images are hierarchically analyzed by region growing and graph editing techniques. Segmented regions are discriptively decided to the final organs by using the semantic informations.

• 한국멀티미디어학회논문지
• /
• 제23권3호
• /
• pp.468-475
• /
• 2020

In this paper, we proposed a three-dimensional visualization system for medical images in augmented reality based on deep learning. In the proposed system, the artificial neural network model performed fully automatic segmentation of the region of lung and pulmonary nodule from chest CT images. After applying the three-dimensional volume rendering method to the segmented images, it was visualized in augmented reality devices. As a result of the experiment, when nodules were present in the region of lung, it could be easily distinguished with the naked eye. Also, the location and shape of the lesions were intuitively confirmed. The evaluation was accomplished by comparing automated segmentation results of the test dataset to the manual segmented image. Through the evaluation of the segmentation model, we obtained the region of lung DSC (Dice Similarity Coefficient) of 98.77%, precision of 98.45%, recall of 99.10%. And the region of pulmonary nodule DSC of 91.88%, precision of 93.05%, recall of 90.94%. If this proposed system will be applied in medical fields such as medical practice and medical education, it is expected that it can contribute to custom organ modeling, lesion analysis, and surgical education and training of patients.

• 대한기계학회논문집A
• /
• 제28권2호
• /
• pp.125-132
• /
• 2004

In this paper, we propose a new two-stage segmentation method for the effective object recognition which uses region-growing algorithm and k-means clustering method. At first, an image is segmented into many small regions via region growing algorithm. And then the segmented small regions are merged in several regions so that the regions of an object may be included in the same region using typical k-means clustering method. This paper also establishes similarity measurement which is useful for object recognition in an image. Similarity is measured by fuzzy system whose input variables are compactness, magnitude of biasness and orientation of biasness of the object image, which are geometrical features of the object. To verify the effectiveness of the proposed two-stage segmentation method and similarity measurement, experiments for object recognition were made and the results show that they are applicable to object recognition under normal circumstance as well as under abnormal circumstance of being.

• JSTS:Journal of Semiconductor Technology and Science
• /
• 제14권4호
• /
• pp.427-435
• /
• 2014

Segmented-channel MOSFETs (SegFETs) can achieve both good performance and variation robustness through the use of $HfO_2$ (a high-k material) to create the shallow trench isolation (STI) region and the very shallow trench isolation (VSTI) region in them. SegFETs with both an HTI region and a VSTI region (i.e., the STI region is filled with $HfO_2$, and the VSTI region is filled with $SiO_2$) can meet the device specifications for high-performance (HP) applications, whereas SegFETs with both an STI region and a VHTI region (i.e., the VSTI region is filled with $HfO_2$, and the STI region is filled with $SiO_2$) are best suited to low-standby power applications. AC analysis shows that the total capacitance of the gate ($C_{gg}$) is strongly affected by the materials in the STI and VSTI regions because of the fringing electric-field effect. This implies that the highest $C_{gg}$ value can be obtained in an HTI/VHTI SegFET. Lastly, the three-dimensional TCAD simulation results with three different random variation sources [e.g., line-edge roughness (LER), random dopant fluctuation (RDF), and work-function variation (WFV)] show that there is no significant dependence on the materials used in the STI or VSTI regions, because of the predominance of the WFV.

• 디지털산업정보학회논문지
• /
• 제9권3호
• /
• pp.13-20
• /
• 2013

In this paper, we propose a multi-scale crack detection method. This method uses decomposition, composition, and shape properties. It is based on morphology algorithm, crack features. We use a morphology operator which extracts patterns of crack. It segments cracks and background using opening and closing operations. Morphology based segmentation is better than existing integration methods using subtraction in detecting a crack it has small width. However, morphology methods using only one structure element could detect only fixed width crack. Thus, we use decomposition and composition methods. We use a decimation method for decomposition. After decomposition and morphology operation, we get edge images given by binary values. Our method calculates values of properties such as the number of pixels and the maximum length of the segmented region. We decide whether the segmented region belongs to cracks according to those data. Experimental results show that our proposed multi-scale crack detection method has better results than those of existing detection methods.

• 대한전자공학회논문지
• /
• 제25권9호
• /
• pp.1091-1101
• /
• 1988

In this paper, a method to segment each character and extract figure from Korean documents is proposed. At first, each character string is extracted by means of iterative horizontal propagation, shrink algorithm and run-length algorithm. Individual character region is extracted by iterative horizontal and vertical manipulation. Next, characters of right pitch are searched. Each character is segmented by the position information. Overlapped character is segmented on the ground of the width of already extracted character. The rest are extracted as special characters of half pitch. Using 9 data input in the form of 840 X 600 from Korean monthly magazine, experiment was simulated. Extraction rate of character is 100%, and that of individual character is 98%. Judging from these results, efficiency on extracting character region and segmenting individual character is proved.