Earticle

Capsule endoscopy is one of the increasingly demanded diagnostic methods among patients in recent years because of its ability to observe small intestine difficulties. It is often conducted for 12 to 14 hours, but significant frames constitute only 10% of whole frames. Thus, it has been designed to automatically acquire significant frames through deep learning. For example, studies to track the position of the capsule (stomach, small intestine, etc.) or to extract lesion-related information (polyps, etc.) have been conducted. However, although grouping or labeling the training images according to similar features can improve the performance of a learning model, various attributes (such as degree of wrinkles, presence of valves, etc.) are not considered in conventional approaches. Therefore, we propose a class-labeling method that can be used to design a learning model by constructing a knowledge model focused on main lesions defined in standard terminologies for capsule endoscopy (minimal standard terminology, capsule endoscopy structured terminology). This method enables the designing of a systematic learning model by labeling detailed classes through differentiation of similar characteristics.

본 논문은 그레이디언트 기반 눈 검출을 위한 비동공 영역 레이블링 방법에 관한 것이다. 제안된 방법은 비동공 영역 레이블링을 이용하여 기존의 그레이디언트 기반 눈 검출 방법보다 개선된 검출 정확도와 연산 속도를 제공함에 그 목적이 있다. 우선 Haar-like feature와 AdaBoost를 이용하여 얼굴 영역을 구한 뒤, 얼굴의 기하학적 특징을 이용하여 좌측과 우측의 눈 탐색 영역을 지정한다. 이후 눈 탐색 영역 내 에지 화소들의 그레이디언트 벡터와 정규 변위 벡터 간의 내적을 누적한 후 최대 누적값의 위치를 좌우 눈의 중심으로 검출한다. 제안된 방법은 동공의 중심은 저명도 평탄 영역에 위치함에 착안하여 비동공 영역으로 추정되는 부분을 내적 누적 연산에서 제외시킴으로써 검출 정확도와 연산 속도를 개선한 것이다. 눈 탐색 영역의 가우시안 필터링된 역 영상을 히스토그램 평활화하여 정규화한 후 임계처리를 통해 동공 후보 영역과 비동공 영역을 구분한다. 하지만 저명도의 안경테나 눈썹, 머리카락과 같은 영역이 후보 영역에 포함될 수 있으므로 비동공 영역 레이블링 기법을 이용해 동공 후보 영역에서 제거하고 전체 눈 탐색 영역이 아닌 후보 영역에서만 그레이디언트 기반 눈 검출을 수행한다. 시뮬레이션 결과에 따르면, 제안된 방법은 기존의 방법[9] 및 개선된 기존의 방법들[10][11]에 비해 우수한 검출 정확도와 연산 속도를 제공하는데, 특히 안경 착용 시 각각 약 51% 및 39.7%, 19% 정도의 연산 성능이 개선되는 장점이 있다.

This paper proposes the labeling method for solving bin-picking problem in robot vision. It has the processing steps such as image thresholding, region labeling, and moment computation. To determine a target object from object, the modified labeling method is used to. The moment concept applied to determine the position and orientation of target object. Finally, some experiment result are illustrated and compared with the results of conventional shrinking algorithm and collision fronts algorithm. The proposed labeling method has reduced processing time.

In the ever-changing landscape of finance, the fusion of artificial intelligence (AI)and pair trading strategies has captured the interest of investors and institutions alike. In the context of supervised machine learning, crafting precise and accurate labels is crucial, as it remains a top priority to empower AI models to surpass traditional pair trading methods. However, prevailing labeling techniques in the financial sector predominantly concentrate on individual assets, posing a challenge in aligning with pair trading strategies. To address this issue, we propose an inventive approach that melds the Triple Barrier Labeling technique with pair trading, optimizing the resultant labels through genetic algorithms. Rigorous backtesting on cryptocurrency datasets illustrates that our proposed labeling method excels over traditional pair trading methods and corresponding buy-and-hold strategies in both profitability and risk control. This pioneering method offers a novel perspective on trading strategies and risk management within the financial domain, laying a robust groundwork for further enhancing the precision and reliability of pair trading strategies utilizing AI models.

Breast cancer classification using ultrasound, while widely employed, faces challenges due to its relatively low predictive value arising from significant overlap in characteristics between benign and malignant lesions, as well as operator-dependency. To alleviate these challenges and reduce dependency on radiologist interpretation, the implementation of automatic breast cancer classification in ultrasound image can be helpful. To deal with this problem, we propose a semi-supervised deep learning framework for breast cancer classification. In the proposed method, we could achieve reasonable performance utilizing less than 50% of the training data for supervised learning in comparison to when we utilized a 100% labeled dataset for training. Though it requires more modification, this methodology may be able to alleviate the time-consuming annotation burden on radiologists by reducing the number of annotation, contributing to a more efficient and effective breast cancer detection process in ultrasound images.

In this study, we present a new isotope-coded carbamidomethylation (iCCM)-based quantitative proteomics, as a complementary strategy for conventional isotope labeling strategies, with providing the simplicity, ease of use, and robustness. In iCCM-based quantification, two proteome samples can be separately isotope-labeled by means of covalently reaction of all cysteinyl residues in proteins with iodoacetamide (IAA) and its isotope (IAA-$^{13}C_2$, $D_2$), denoted as CM and iCCM, respectively, leading to a mass shift of all cysteinyl residues to be + 4 Da. To evaluate iCCM-based isotope labeling in proteomic quantification, 6 protein standards (i.e., bovine serum albumin, serotransferrin, lysozyme, beta-lactoglobulin, beta-galactosidase, and alpha-lactalbumin) isotopically labeled with IAA and its isotope, mixed equally, and followed by proteolytic digestion. The resulting CM-/iCCM-labeled peptide mixtures were analyzed using a nLC-ESI-FT orbitrap-MS/MS. From our experimental results, we found that the efficiency of iCCM-based quantification is more superior to that of mTRAQ, as a conventional nonisobaric labeling method, in which both of a number of identified peptides from 6 protein standards and the less quantitative variations in the relative abundance ratios of heavy-/light-labeled corresponding peptide pairs. Finally, we applied the developed iCCM-based quantitative method to lung cancer serum proteome in order to evaluate the potential in biomarker discovery study.

The molecular imaging techniques allow monitoring of the transplanted cells in the same individuals over time, from early localization to the survival, migration, and differentiation. Generally, there are two methods of stem cell labeling: direct and indirect labeling methods. The direct labeling method introduces a labeling agent into the cell, which is stably incorporated or attached to the cells prior to transplantation. Direct labeling of cells with radionuclides is a simple method with relatively fewer adverse events related to genetic responses. However, it can only allow short-term distribution of transplanted cells because of the decreasing imaging signal with radiodecay, according to the physical half-lives, or the signal becomes more diffuse with cell division and dispersion. The indirect labeling method is based on the expression of a reporter gene transduced into the cell before transplantation, which is then visualized upon the injection of an appropriate probe or substrate. In this review, various imaging strategies to monitor the survival and behavior change of transplanted stem cells are covered. Taking these new approaches together, the direct and indirect labeling methods may provide new insights on the roles of in vivo stem cell monitoring, from bench to bedside.

A $^{18}F$-labeled glucose analog, 4-[(2-[$^{18}F$]fluoroethyl)-1-($\beta$-D-glucopyranosyl)]-1H-1,2,3-triazole ([$^{18}F$]1), was synthesized using a click labeling method and evaluated in vitro for its cellular transportation via glucose transporter (Glut-1) and its potential as a hexokinase substrate. The click labeling method was superior to conventional labeling method, due to a higher decay-corrected radiochemical yield (30% vs. 21%), higher specific activity ($59.9\;GBq/{\mu}mol$ vs. $23.5\;GBq/{\mu}mol$), and shorter synthesis time (75-80 min vs. 95-100 min). In vitro evaluation demonstrated that [$^{18}F$]1 does not act as a hexokinase substrate and has low and non-specific uptake by SNU-C5 cells. These results suggest that click chemistry offers a rapid and efficient radiolabeling method which does not require the protection of functional groups, although a triazole moiety at C1 of [$^{18}F$]1 is incompatible for hexokinase phosphorylation and facilitative diffusion via Glut-1.

Treponema denticola is a gram-negative anaerobe that can cause periodontal disease. The adhesion of this bacterium to host tissues is considered to be the primary event in the colonization and infection of a host. Fibrinogen is generally found in damaged tissues resulting from periodontitis. The binding ability of T. denticola to fibrinogen may therefore be an important virulence factor in inducing periodontal diseases. It has been reported recently that oral spirochetes can be labeled with fluorescent fatty acids and we speculated that this labeling method could be used in an oral spirochete binding assay. The binding of several different strains of T. denticola to immobilized human fibrinogen was therefore tested using the fluorescent fatty acid labeling method. In the case of immobilized fibrinogen, the T. denticola ATCC 35405 strain showed saturable binding to immobilized fibrinogen. Indeed, all four different T. denticola strains tested in this experiment, T. denticola ATCC 35405, T. denticola ATCC 33520, T. denticola ATCC 35404 and T. denticola OTK showed binding to fibrinogen. The fluorescent fatty acid labeling method thus shows utility in binding assays for T. denticola, different strains of which can generally bind to immobilized fibrinogen.

Treponema denticola is a gram-negative anaerobe that can cause periodontal disease. The adhesion of this bacterium to host tissues is considered to be the primary event in the colonization and infection of a host. Fibrinogen is generally found in damaged tissues resulting from periodontitis. The binding ability of T. denticola to fibrinogen may therefore be an important virulence factor in inducing periodontal diseases. It has been reported recently that oral spirochetes can be labeled with fluorescent fatty acids and we speculated that this labeling method could be used in an oral spirochete binding assay. The binding of several different strains of T. denticola to immobilized human fibrinogen was therefore tested using the fluorescent fatty acid labeling method. In the case of immobilized fibrinogen, the T. denticola ATCC 35405 strain showed saturable binding to immobilized fibrinogen. Indeed, all four different T. denticola strains tested in this experiment, T. denticola ATCC 35405, T. denticola ATCC 33520, T. denticola ATCC 35404 and T. denticola OTK showed binding to fibrinogen. The fluorescent fatty acid labeling method thus shows utility in binding assays for T. denticola, different strains of which can generally bind to immobilized fibrinogen.

Objective: The objective of a conservation program is to maintain maximum genetic diversity and preserve the viability of a breed. However, the efficiency of a program is influenced by the ability to accurately measure and predict genetic diversity. Methods: To examine this question, we conducted a simulation in which common measures (i.e. heterozygosity) and novel measures (identity-by-descent probabilities and parental genomic components) were used to estimate genetic diversity within a conserved population using double-labeled single nucleotide polymorphism markers. Results: The results showed that the accuracy and sensitivity of identity-by-state probabilities and heterozygosity were close to identity by descent (IBD) probabilities, which reflect the true genetic diversity. Expected heterozygosity most closely aligned with IBD. All common measures suggested that practices used in the current Chinese pig conservation program result in a ~5% loss in genetic diversity every 10 generations. Parental genomic components were also analyzed to monitor real-time changes in genomic components for each male and female ancestor. The analysis showed that ~7.5% of male families and ~30% of female families were lost every 5 generations. After 50 generations of simulated conservation, 4 male families lost ~50% of their initial genomic components, and the genomic components for 24.8% of the female families were lost entirely. Conclusion: In summary, compared with the true genetic diversity value obtained using double-labeled markers, expected heterozygosity appears to be the optimal indicator. Parental genomic components analysis provides a more detailed picture of genetic diversity and can be used to guide conservation management practices.

A new graph-theoretic clustering method based on relaxation labeling is proposed. More accurate classification is achieved by using relaxation labeling which take advantage of contextual information. Moreover, threshold selection is not needed in this method.

This paper proposes a new automatic detection method to inspect mura defects on display film surface using morphological image processing and labeling. This automatic detection method for mura defects on display films comprises 3 phases of preprocessing with morphological image processing, Gabor filtering, and labeling. Since distorted results could be obtained with the presence of non-uniform illumination, preprocessing step reduces illumination components using morphological image processing. In Gabor filtering, mura images are created with binary coded mura components using Gabor filters. Subsequently, labeling is a final phase of finding the mura defect area using the difference between large mura defects and values in the periphery. To evaluate the accuracy of the proposed detection method, detection rate was assessed by applying the method in 200 display film samples. As a result, the detection rate was high at about 95.5%. Moreover, the study was able to acquire reliable results using the Semu index for luminance mura in image quality inspection.

인터넷을 통해 전송되는 트래픽의 양이 점점 더 많아지고 있고, 이에 따라 트래픽의 양을 알아내는 것이 중요해지고 있다. 트래픽을 측정하는 기법에 대한 많은 연구가 있었으며, 주로 적은 양의 메모리를 사용해서 측정의 정확도를 높이는 방향으로 연구가 진행되고 있다. 이 논문에서는 기존 기법들에서는 제공하지 않는 플로우 라벨링 기법을 제안한다. 이 기법을 통해서 관리자는 특정 값 이상의 트래픽을 생성한 플로우의 라벨을 실시간으로 알아낼 수 있으며, 기존의 기법이 가지는 카운팅의 상한선을 확장시킬 수 있다. 가장 최근에 발표된 CSM (Counter Sharing Method)에 이 기법을 적용하고, CAIDA 데이터셋을 이용해서 성능을 분석해 본다.

As a greater amount of traffic have been generated on the Internet, it becomes more important to know the size of each flow. Many research studies have been conducted on the traffic measurement, and mostly they have focused on how to increase the measurement accuracy with a limited amount of memory. In this paper, we propose an explicit flow labeling technique that can be used to find out the names of the top flows and to increase the counting upper bound of the existing scheme. The labeling technique is applied to CSM(Counter Sharing Method), the most recent traffic measurement algorithm, and the performance is evaluated using the CAIDA dataset.

수중 환경에서의 객체 탐지는 물체의 겹침, 노이즈, 조명 변화 등으로 인해 어려움이 있으며, 특히 넙치와 같은 수중 생물의 경우 수동 라벨링에 많은 시간과 노력이 필요하다. 본 논문에서는 이러한 문제를 해결하기 위해 영상처리 기법을 활용하여 넙치의 라벨링을 자동화하고 학습을 위한 데이터셋의 크기를 증가 방법을 제안한다. 제안한 방법은 모폴로지 연산과 외곽선 검출 방법을 적용하여 노이즈를 제거하고 객체의 테두리를 정확하게 추출하였으며, YOLOv8 모델을 사용하여 데이터셋의 크기와 학습 방법에 따른 성능 변화를 분석했다. 그 결과, 자동화된 이미지 전처리를 통해 데이터셋을 확대함으로써 겹침이 발생한 객체 탐지에서도 향상된 성능을 얻을 수 있었고, 수동 라벨링에 비해 시간적 효율성이 높음을 확인하였다.

Detecting objects in underwater environments is challenging due to factors such as object overlap, noise, and lighting variations. In particular, labeling underwater organisms like Paralichthys olivaceus requires significant time and effort when done manually. To address these challenges, this paper proposes a method to automate the labeling of Paralichthys olivaceus using image processing techniques and to increase the size of the dataset for training. The proposed method applies morphological operations and contour detection to remove noise and accurately extract object boundaries. Additionally, the Y0L0v8 model is used to analyze performance variations based on dataset size and training methods. The results demonstrate that expanding the dataset through automated image preprocessing improves object detection performance, even in cases of object overlap, and proves to be more time-efficient compared to manual labeling.

캡슐내시경 검사는 일반내시경으로는 관찰하기 힘든 소장 기관을 관찰할 수 있어 최근 환자들 사이에서 수요가 늘고 있는 검사 방법 중 하나이다. 이와 같은 캡슐내시경으로부터 병변에 대한 의료 정보가 획득될 수 있는데, 최근에는 캡슐내시경 영상의 학습을 통해 이를 자동으로 획득하려는 시도들이 이루어지고 있다. 예를 들면, 캡슐의 위치를 추적하기 위해 위장관의 개략적인 위치(위, 소장 등)를 파악하거나, 캡슐내시경 영상으로부터 관찰될 수 있는 병변(폴립 등)을 검출하기 위해 영상의 학습이 수행되고 있는 상황이다. 그러나 캡슐내시경의 방대한 영상 프레임 중에서 병변에 대한 영상은 극히 일부분이기 때문에, 기존 학습 영상의 클래스(레이블)는 다양한 병변에 대한 정의나 영상에서 확인될 수 있는 구체적인 속성이 고려되지 않는다. 따라서 본 논문에서는 캡슐내시경 관련 표준(MST, CEST)에서 정의하고 있는 주요 병변 정보에 대한 색상 유사도 분석을 통해, 출력층에서 활용될 수 있는 클래스 레이블링 기법을 제안한다. 제안하는 기법은 유사한 특성을 보이는 영상의 구분을 통해 세부적인 클래스 레이블링을 수행하여 체계적인 학습 모델의 설계를 가능케한다.

캡슐내시경 검사는 일반내시경으로는 관찰하기 힘든 소장 기관을 관찰할 수 있어 최근 환자들 사이에서 수요가 늘고 있는 검사 방법 중 하나이다. 이와 같은 캡슐내시경으로부터 병변에 대한 의료 정보가 획득될 수 있는데, 최근에는 캡슐내시경 영상의 학습을 통해 이를 자동으로 획득하려는 시도들이 이루어지고 있다. 예를 들면, 캡슐의 위치를 추적하기 위해 위장관의 개략적인 위치(위, 소장 등)를 파악하거나, 캡슐내시경 영상으로부터 관찰될 수 있는 병변(폴립 등)을 검출하기 위해 영상의 학습이 수행되고 있는 상황이다. 그러나 캡슐내시경의 방대한 영상 프레임 중에서 병변에 대한 영상은 극히 일부분이기 때문에, 기존 학습 영상의 클래스(레이블)는 다양한 병변에 대한 정의나 영상에서 확인될 수 있는 구체적인 속성이 고려되지 않는다. 따라서 본 논문에서는 캡슐내시경 관련 표준(MST, CEST)에서 정의하고 있는 주요 병변 정보에 대한 색상 유사도 분석을 통해, 출력층에서 활용될 수 있는 클래스 레이블링 기법을 제안한다. 제안하는 기법은 유사한 특성을 보이는 영상의 구분을 통해 세부적인 클래스 레이블링을 수행하여 체계적인 학습 모델의 설계를 가능케한다.

세라믹 소재 영상에서 결함 영역이 다른 영역보다 명암도가 밝게 나타나는 정보를 이용하여 ROI 영역을 추출한다. 추출된 ROI 영역에서 Blurring 기법을 적용하여 미세 잡음을 제거한다. 미세 잡음이 제거된 ROI 영역에서 Median Filter기법을 적용하여 임펄스 잡음을 제거한다. 임펄스 잡음이 제거된 영역에서 Prewit Mask을 적용하여 수평과 수직 에지를 검출하고 검출된 에지에 윤곽선 추적 기법을 적용하여 결함 영역의 경계를 보정한다. 보정된 영상에서 Blob Labeling 기법을 적용하여 최종적으로 결함 영역을 추출한다. 제안된 방법을 8mm와 10mm 세라믹 소재 영상을 대상으로 실험한 결과, 기존의 결함 검출 방법보다 제안된 검출 방법의 검출 성능이 개선된 것을 확인하였다.