• 한국소음진동공학회논문집
• /
• 제22권12호
• /
• pp.1237-1242
• /
• 2012

No matter how low it might be, noise from home appliances in indoor environment cause displeasure for residents as it is produced nearby. Electronic goods including vacuum cleaner and washing machine generate high frequency noise, which interrupts TV-watching or conversations and cause noises between floors in apartment houses. In particular, refrigerators make constant noise late at night, hampering the study for exams and causing sleeplessness. Korea, despite the government's efforts to promote low-noise goods based on the noise test by corporation, there is a lack of standardization in the test method, test condition and measuring equipments. This is a major reason that lowers the reliability of low-noise products. At present, low-noise home appliances are certified with eco labeling in the nation but, only 7 brands of refrigerators, 4 brands of washing machines and 97 brands of air-conditioners obtained certification as of august, 2011. In addition, none of Kimchi-refrigerator and vacuum cleaner brands were approved as low-noise home appliances. This shows Korea's relatively underdeveloped market for low-noise products, which, in part, is because of the difficulties companies face in being certified with eco labeling as they should satisfy various requirements such as power-saving and eco-friendly design besides low-noise features. As a result, low-noise labeling for home appliances should be established for the revised noise and vibration management law and the study was carried out to establish low-noise labeling system for major noise sources including vacuum cleaner, washing machine, air-conditioner and refrigerator.

• 한국통신학회논문지
• /
• 제36권9C호
• /
• pp.555-563
• /
• 2011

본 논문에서는 움직임 추정 (motion estimation, ME), 컬러 라벨링(labeling) 그리고 Non-local mean 필터를 이용하여 2D 영상을 3D 업체 영상으로 변환하는 기법을 제안한다. 제안하는 기법에서는 먼저 프레임 간의 움직임을 추정하여 객체의 움직임 벡터를 추출하고 주어진 영상에 대해 컬러 라벨링 작업을 수행하여 영상을 분리한다. 움직임 추정 결과와 컬러 라벨링 결과를 비교 분석하여 영상내의 객체를 추출하고 추출된 객체를 이동하여 우 영상을 생성하게 되는데 이때 우 영상을 생성하는 과정에서 채워지지 않은 가려짐 영역이 발생하며 전체 화소간의 상관도를 고려하는 Non-local mean 필터를 사용하여 보상한다. 이후 원본 영상인 좌 영상과 생성된 우 영상으로 비윌 주사하여 최종 3D 업체 영상을 재현한다. 실험 결과를 통해 제안된 기법으로 생성된 3D 업체 영상에서 객체위주의 안정된 업체 변환이 수행되는 것을 확인할 수 있었다.

• 대한지역사회영양학회지
• /
• 제9권5호
• /
• pp.636-643
• /
• 2004

The purpose of this study was to investigate school dietitians' perception on nutrition labelling of processed and packaged foods in the Incheon area. This study was carried out through a self-administered questionnaire and the subjects were 203 school dietitians. The results were summarized as follows. Most of the subjects checked nutrition labeling of processed and packaged foods when they purchased these foods for school lunch. However, the main reason for their checking nutrition labeling of processed and packaged foods was not for nutrition but for food safety. There were significant differences in the subjects' understanding and trust on nutrition labeling of processed and packaged foods by employment school. The scores of the subjects' satisfaction, understanding and trust on nutrition labeling of processed and packaged foods were not high. There were significant differences in the subjects' perception and concern about content information on nutrition labeling of domestic processed and packaged foods by employment school. Most of the subjects answered that it is almost necessary or very necessary to implement nutrition labelling of processed and packaged foods. Also their most preferred method of nutrition labelling was to label nutrient content. As the most important nutrient and functional component necessary for nutrition labelling, they answered energy, mineral such as calcium and iron, protein, total fat, cholesterol, vitamin, sodium and dietary fiber in order. As the major prerequisite task for the implementation of nutrition labelling, they answered formulation of nutrition labelling-related rules, and education and motivation for consumer, Therefore, it is necessary to activate the mandatory nutrition labelling of the processed and packaged foods and to prepare its consumer education program for school dietitian.

• 대한토목학회논문집
• /
• 제39권5호
• /
• pp.631-636
• /
• 2019

최근 많은 연구자들이 대규모 현장에 투입된 건설자원의 유형과 위치를 자동 파악하는 영상분석기술을 활발히 개발하고 있다. 하지만 기존의 방법들은 인식하고자 하는 건설 물체(작업자, 중장비, 자재 등)를 학습용 이미지 데이터에 표시하는 Labeling 작업을 요구하고 이에 불필요한 시간과 노력이 낭비된다는 한계가 있다. 이러한 한계를 보완하기 위해서 본 연구는 액티브 러닝을 활용한 영상기반 건설현장 물체 자동 인식 프레임 워크를 제안함을 목표로 한다. 개발 프레임워크 검증을 목적으로 건설분야 Benchmark 데이터셋을 이용하여 실제 실험을 진행하였다. 그 결과, 액티브 러닝을 통해 학습한 모델은 다양한 특성을 지닌 건설물체를 성공적으로 인식할 수 있었고, 기존의 학습 DB 구축 방식과 비교할 때 더 적은 데이터 수와 반복학습 횟수로도 높은 성능을 가지는 영상분석모델을 개발할 수 있었다. 결과적으로 기존에 요구되던 학습 DB 구축을 위한 Labeling 작업을 줄일 뿐만 아니라 총 시간과 비용을 최소화할 수 있다.

• 한국정보통신학회논문지
• /
• 제9권7호
• /
• pp.1575-1581
• /
• 2005

최근에 콘크리트 구조물에 발생한 균열의 영향에 대한 관심이 커져 과거와 다르게 발생한 균열을 방치하기보다 보수를 통해 그 영향을 최소화시키고자 노력하고 있다. 본 논문에서는 콘크리트 표면 균열의 효율적인 측정을 위하여 점검자의 육안조사를 통한 수작업 대신 컴퓨터가 균열의 길이, 방향, 폭등의 특징을 자동으로 처리하는 화상처리 기법을 제안한다. 정확한 균열의 특징 추출을 위하여 빛의 영향을 보정할 수 있는 모폴로지 기법인 채움(Closing)연산을 적용하여 영상의 빛의 영향을 보정하고 Sobel 마스크를 적용하여 균열영상의 에지를 추출한 후, 반복 이진화를 적용하여 균열 영상을 이진화한다 그리고 이진화된 영상에서 잡음을 제거하기 위해 3${\times}$3 마스크를 이용하여 인접화소의 평균을 이용한 잡음 제거와 Glassfire Labeling 알고리즘을 이용한 잡음 제거를 각각 수행하여 균열의 에지를 추출하게 된다. 그리고 본 논문에서는 추출된 균열의 에지를 이용하여 균열의 특정부분의 길이, 방향, 폭을 자동으로 측정하는 표면 균열의 자동 처리 기법을 제안한다. 실험 결과에서는 제시된 방법이 균열의 검출에 효율적임을 나타내었고 검시자의 주관이 개입될 수 있는 가능성을 제거할 수 있는 가능성을 나타내었다.

• 대한의용생체공학회:의공학회지
• /
• 제28권4호
• /
• pp.549-555
• /
• 2007

In order to study cortical properties in human, it is necessary to obtain an accurate and explicit representation of the cortical surface in individual subjects. Among many approaches, surface-based method that reconstructs a 3-D model from contour lines on cross-section images is widely used. In general, however, medical brain imaging has some problems such as the complexity of the images, non-linear gain artifacts and so on. Due these limitations, therefore, extracting anatomical structures from imaging data is very a complicated and time-consuming task. In this paper, we present an improved method for extracting contour lines of cortical surface from magnetic resonance images that simplifies procedures of a conventional method. The conventional method obtains contour lines through thinning and chain code process. On the other hand, the proposed method can extract contour lines from comparison between boundary data and labeling image without supplementary processes. The usefulness of the proposed method has been verified using brain image.

• Fisheries and Aquatic Sciences
• /
• 제4권1호
• /
• pp.10-17
• /
• 2001

One serious problem among the troubles to identify objects in an optical microscopic image is contour background due to non-uniform light source and various transparency of samples. To solve this problem, this paper proposed an elimination method of the contour background and compensation technique as follows. First, Otsu's optimal thresholding method extracts pixels representing background. Second, bilinear interpolation finds non-deterministic background pixels among the sampled pixels. Third, the 2D cubic fitting method composes surface function from pivoted background pixels. Fourth, reconstruction procedure makes a contour image from the surface function. Finally, elimination procedure subtracts the approximated background from the original image. To prove the effectiveness of the proposed algorithm, this algorithm is applied to the yeast Zygosaccharomyces rouxii and ammonia-oxidizing bacteria Acinetobacter sp. Labeling by this proposed method can remove some noise and is more exact than labeling by only Otsu's method. Futhermore, we show that it is more effective for the reduction of noise.

• 인터넷정보학회논문지
• /
• 제18권1호
• /
• pp.43-50
• /
• 2017

본 논문에서는 한 장의 실내이미지에서 내벽구조 검출을 위한 개선된 소실점 검출방법과 세그먼트 레이블링 방법을 제안한다. AR 기술 수요의 증가로 이미지로부터 건축물의 구조를 인식하는 것과 관련된 연구가 많이 이루어 지고 있다. 그러나 폐색을 발생시키는 객체들이 많은 실내 이미지에서 실내 내부 구조를 인식하는 것은 여전히 어려운 문제이다. 소실점 검출 방법을 개선하기 위하여 육면체를 이루는 실내 내벽들 사이의 직교성을 이용하는 방법을 제안하였다. 또한 실내 이미지 내의 세그먼트들을 레이블링 하기 위하여 슈퍼픽셀 기반의 군집화 방법과 트리기반 학습기를 통한 레이블링 방법을 제안하였다. 마지막으로 실험 결과에서 제안한 방법들에 의하여 실내 구조 검출 결과가 개선됨을 보였다.

• Animal cells and systems
• /
• 제3권1호
• /
• pp.73-78
• /
• 1999

Although there are several methods for the preparation of mitochondrial DNA (mtDNA) from mammalian tissues, most are relatively long ultracentrifugation or manipulations by a small-scale method. We escribed a rapid method for large-scale extraction of mtDNA from human placental and horse liver tissues. The method is based on the preparation and homogenization of tissues, urification of crude mitochondria by differential centrifugations and isolation of mtDNA by alkaline Iysis. It was improved from Pre-existing methods by replacing some steps with simpler ones and discarding many others. This method gives a high yield of pure mtDNA(approximately 1-5mg from one placenta; ca. 400-600 g wet weight), depending on its sources (fresh tissue gave better results than frozen one). The resulting mtDNA indicated that this method can yield mtDNA in sufficient purity and quantity to identify the direct restriction analysis on agarose gel, random-primed labeling as a probe, and end labeling. Therefore, the method is ideal for obtaining good mtDNA samples to conduct routine restriction fragment length polymorphism (RFLP) analyses of natural populations for genetic studies.

• 대한방사성의약품학회지
• /
• 제1권1호
• /
• pp.15-22
• /
• 2015

$^{18}F$-labeling reaction of bioactive molecule via click chemistry is widely used to produce $^{18}F$-labeled radiotracer in the field of radiopharmaceutical science and molecular imaging. In particular, bioorthogonal strain-promoted alkyne-azide cycloaddition (SPAAC) reaction has received much attention as an alternative ligation method for radiolabeling bioactive molecules such as peptides, DNA, proteins as well as nanoparticles. Moreover, SPAAC based pretargeting method could provide tumor images successfully on positron emission tomography system using nanoparticle such as mesoporous silica nanoparticles.