• 전자공학회논문지CI
• /
• 제45권3호
• /
• pp.97-102
• /
• 2008

본 논문은 컬러 질의 영상의 효과적인 검출을 위해 공간 컬러모델 및 특징점 정합 방법을 이용한 객체 기반 영상 검색 방법을 제안한다. 제안하는 방법은 선행 연구 되었던 컬러 히스토그램 방법의 단점을 극복하고, 데이터베이스 영상과 질의 영상의 컬러 유사도를 사용자 조작 없이 실시간 분할 검출한다. 이를 위해 HMMD 모델과 러프 집합 이론을 이용하였다. 여기서 질의 영상의 검출을 위해 질의 영상과 데이터베이스 영상 간의 색상 유사도를 비교하여 관심 영역을 선택하고, 관심 영역에서 SIFT 정합 방법을 이용하여 검색한다. 실험 결과, 본 논문에서 제안하는 방법이 기존 방법보다 우수한 검출율을 보임을 확인하였다.

• 대한공간정보학회지
• /
• 제13권4호
• /
• pp.33-38
• /
• 2005

본 연구는 항공레이저스캐너의 스캔라인 특성을 활용하여 건물 포인트를 효율적으로 분리하는 것을 목표로 한다. 포인트 간의 고도 유사성 및 인접성을 기준으로 포인트들을 분류하였으며, 분류 대상 클래스의 탐색 범위를 소수의 스캔라인으로 제한함으로써 분류가 진행됨에 따라 분류 속도가 저하되는 현상을 방지하였다 또한 건물의 형태 및 스캔라인의 특성으로 인해 동일 개체가 두 개 이상의 클래스로 분리되는 현상을 감지하고 하나의 클래스로 통합하는 기능도 구현하였다. 결과적으로 개별 건물, 옥탑과 같은 부속 건물, 비건물 포인트를 동시에 분리할 수 있었다.

• Asian Journal of Business Environment
• /
• 제7권1호
• /
• pp.25-29
• /
• 2017

Purpose - In this paper, we categorize and segment the 28 national universities in South Korea and measure the degree of dissimilarity (or similarity) between pairs of ones by using dissimilarity distance matrix and cluster analysis, respectively, based on the seven quantitative evaluation of educational conditions (percentage of small-scale courses, percentage of lecture by the faculty, collection of books per student, material purchase per student, percentage of building capacity, percentage of real estate capacity and rate of accommodation) in 2015. In addition, multidimensional scaling (MDS) techniques can obtain visual representation for exploring patterns of proximities among 28 national universities based on seven attributes of educational conditions. Research design, data, and methodology - This work is carried out by the 2015 Announcement of University Information, which is provided by Ministry of Education in South Korea and utilized by multivariate analyses with CLUSTER, PROXIMITIES and ALSCAL modules in IBM SPSS 23.0. Results - We make certain that 28 national universities can be categorized into five clusters which have similar traits by applying two-stage cluster analysis. MDS is utilized to perform positioning of grouped places of cluster and 28 national universities joining every cluster. Conclusions - Both types and traits of each national university can be relatively assessed and practically utilized for each university competitiveness based on underlying results.

• The Plant Pathology Journal
• /
• 제30권2호
• /
• pp.151-158
• /
• 2014

Three double-stranded RNAs (dsRNAs), approximately 1.85, 1.65 and 1.27 kb in size, were detected in an isolate of Cytospora sacchari from Iran. Partial nucleotide sequence revealed a 1,284 bp segment containing one ORF that potentially encodes a 405 aa protein. This protein contains conserved motifs related to RNA dependent RNA polymerases (RdRp) that showed similarity to RdRps of partitiviruses. The results indicate that these dsRNAs represent a novel Partitivirus that we tentatively designate Cytospora sacchari partitivirus (CsPV). Treatment of the fungal strain by cyclohexamide and also hyphal tip culture had no effect on removing the putative virus. Phylogenetic analysis of putative RdRp of CsPV and other partitiviruses places CsPV as a member of the genus Partitivirus in the family Partitiviridae, and clustering with Aspergillus ochraceous virus 1.

• 한국식품조리과학회지
• /
• 제6권1호
• /
• pp.13-19
• /
• 1990

Cowpea, mung bean, Kidney bean and red bean are simular properties. In order to elucidate the similarity among these four starches, some physicochemical properties of starches were compared. Water binding capacity of kidney bean and red bean (199%) starches are higher than mung bean and cowpea. The solubility, swlling power and optical transmittance of the four starches showed a similar pattern, but kidney bean and red bean starches had a lower swelling power than cowpea starches. Cowpea, mung bean, kidney bean and red bean starches had the blue value of 0.41, 0.47, 0.42 and 0.50, the alkali content of 8.4, 8.0, 4.13, 4.13, the amylose content of amylose of 30,000, 29,268, 52, 173 and 33, 611 and glucose unit per segment of amylopectin of 27.6, 26.8, 18.35 and 12.9 respectively. The results of X-ray diffraction studies showed A pattern for four starches.

• Journal of Ginseng Research
• /
• 제32권2호
• /
• pp.114-119
• /
• 2008

Terpene synthase plays a key role in biosynthesis of triterpene saponins (ginsenosides) and is intermediate in the biosynthesis of a number of secondary metabolites. A terpene synthase (PgTPS) cDNA was isolated and characterized from the root of Panax ginseng c.A. Meyer. The deduced amino acid sequence of PgTPS showed a similarity with A. deliciosa (AAX16121) 61%, V. vinifera (AAS66357) 61%, L. hirsutum (AAG41891) 55%, M. truncatula (AAV36464) 52%. And the segment of a terpene synthase gene was amplified by reverse transcriptase-polymerase chain reaction (RTPCR). We studied expression of terpene synthase under stressful conditions like chilling, salt, UV, and heavy metal stress treatment. Expression of PgTPS was increased gradually after exposure to stresses such as chilling, salt, and UV illumination. But its transcription seems to be reduced by cadmium and copper treatment.

• 한국약용작물학회지
• /
• 제16권4호
• /
• pp.261-267
• /
• 2008

Malate dehydrogenase is a ubiquitous enzyme in plants, involving in a range of metabolic processes depending on its subcellular location. A malate dehydrogenase (PgMDH) cDNA was isolated and characterized from the root of Panax ginseng C. A. Meyer. The deduced amino acid sequence of PgMDH showed high similarity with the NAD-dependent mitochondrial malate dehydrogenase from Glycinemax (P17783), Eucalyptus gunnii (P46487), and Lycopersicon esculentum (AAU29198). And the segment of a malate dehydrogenase gene was amplified through RT-PCR. The expression of PgMDH was increased after treatments of chilling, salt, UV, cadmium or copper treatment.

• 한국정보통신학회논문지
• /
• 제14권8호
• /
• pp.1819-1826
• /
• 2010

본 논문은 어류영상 검색을 위해 어류영상에서 어류객체를 추출하기 위한 영상분할 알고리즘을 제안하고 있다. 명암 유사도를 이용한 기존 알고리즘은 객체와 배경의 명암이 유사한 경계 영역에서 잘못된 영상분할 결과를 초래한다. 제안된 알고리즘은 대비가 약한 경계영역에 대응하기 위해 강화된 에지와 적응적 블록단위의 임계값을 사용하고, 대비가 없는 경계 영역에서 침식 혹은 단절된 객체를 개선하기 위해 가상 객체를 사용하고 있다. 모의실험 결과는 시각적으로 좋은 어류객체를 추출하는 비율이 기존 알고리즘에서는 90% 이하인 반면 제안된 알고리즘에서는 97.7%인 것을 보여주고 있다.

• 디지털산업정보학회논문지
• /
• 제14권4호
• /
• pp.69-77
• /
• 2018

The malignant melanoma accounts for about 1 to 3% of the total malignant tumor in the West, especially in the US, it is a disease that causes more than 9,000 deaths each year. Generally, skin lesions are difficult to detect the features through photography. In this paper, we propose a computer-aided diagnosis algorithm based on deep learning for classification of malignant melanoma and benign skin tumor in RGB channel skin images. The proposed deep learning model configures the tumor lesion segmentation model and a classification model of malignant melanoma. First, U-Net was used to segment a skin lesion area in the dermoscopic image. We could implement algorithms to classify malignant melanoma and benign tumor using skin lesion image and results of expert's labeling in ResNet. The U-Net model obtained a dice similarity coefficient of 83.45% compared with results of expert's labeling. The classification accuracy of malignant melanoma obtained the 83.06%. As the result, it is expected that the proposed artificial intelligence algorithm will utilize as a computer-aided diagnosis algorithm and help to detect malignant melanoma at an early stage.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제18권1호
• /
• pp.15-29
• /
• 2024

A fetoscope is an optical endoscope, which is often applied in fetoscopic laser photocoagulation to treat twin-to-twin transfusion syndrome. In an operation, the clinician needs to observe the abnormal placental vessels through the endoscope, so as to guide the operation. However, low-quality imaging and narrow field of view of the fetoscope increase the difficulty of the operation. Introducing an accurate placental vessel segmentation of fetoscopic images can assist the fetoscopic laser photocoagulation and help identify the abnormal vessels. This study proposes a method to solve the above problems. A novel encoder-decoder network with a dual-path structure is proposed to segment the placental vessels in fetoscopic images. In particular, we introduce a channel attention mechanism and a continuous convolution structure to obtain multi-scale features with their weights. Moreover, a switching connection is inserted between the corresponding blocks of the two paths to strengthen their relationship. According to the results of a set of blood vessel segmentation experiments conducted on a public fetoscopic image dataset, our method has achieved higher scores than the current mainstream segmentation methods, raising the dice similarity coefficient, intersection over union, and pixel accuracy by 5.80%, 8.39% and 0.62%, respectively.