• 한국멀티미디어학회논문지
• /
• 제19권8호
• /
• pp.1248-1259
• /
• 2016

Classification methods based on dual energy X-ray absorptiometry, ultrasonic waves, and quantitative computed tomography have been proposed. Also, a classification method based on machine learning with bone mineral density and structural indicators extracted from the CT images has been proposed. We propose a method which enhances the performance of existing classification method based on bone mineral density and structural indicators by extending structural indicators and using principal component analysis. Experimental result shows that the proposed method in this paper improves the correctness of osteoporosis classification 2.8% with extended structural indicators only and 4.8% with both extended structural indicators and principal component analysis. In addition, this paper proposes a method of automatic phantom analysis needed to convert the CT values to BMD values. While existing method requires manual operation to mark the bone region within the phantom, the proposed method detects the bone region automatically by detecting circles in the CT image. The proposed method and the existing method gave the same conversion formula for converting CT value to bone mineral density.

• Composites Research
• /
• 제34권2호
• /
• pp.96-100
• /
• 2021

The lightweight industry continuously demands reliable near-net-shape fabrication where the preform just out-of-machine is close to the final shape. In this study, different half-finished preforms are made π-beams. Then the preforms are unfolded to make a 3D shape with integrated structure of fibers, providing easier handling in the further processing of composites. Several 3D textile preforms are made using weaving technique and are examined after resin infusion for mechanical properties such as inter-laminar shear strength, compressive strength and tensile strength. Considering that the time and labor are important parameters in modern production, 3D weaving technique reduces the manufacturing steps and therefore the costs, such as hand-lay up of textile layers, cutting, and converting into preform shape. Hence this 3D weaving technique offers many possibilities for new applications with efficient composite production.

• ETRI Journal
• /
• 제42권6호
• /
• pp.899-911
• /
• 2020

The data in tables are accurate and rich in information, which facilitates the performance of information extraction and question answering (QA) tasks. TableQA, which is based on tables, solves problems by understanding the table structure and searching for answers to questions. In this paper, we introduce both novice and intermediate Korean TableQA tasks that involve deducing the answer to a question from structured tabular data and using it to build a question answering pair. To solve Korean TableQA tasks, we use S3-NET, which has shown a good performance in machine reading comprehension (MRC), and propose a method of converting structured tabular data into a record format suitable for MRC. Our experimental results show that the proposed method outperforms a baseline in both the novice task (exact match (EM) 96.48% and F1 97.06%) and intermediate task (EM 99.30% and F1 99.55%).

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권2호
• /
• pp.962-977
• /
• 2019

The human voice is a convenient method of information transfer between different objects such as between men, men and machine, between machines. The development of information and communication technology, the voice has been able to transfer farther than before. The way to communicate, it is to convert the voice to another form, transmit it, and then reconvert it back to sound. In such a communication process, a vocoder is a method of converting and re-converting a voice and sound. The CELP (Code-Excited Linear Prediction) type vocoder, one of the voice codecs, is adapted as a standard codec since it provides high quality sound even though its transmission speed is relatively low. The EVRC (Enhanced Variable Rate CODEC) and QCELP (Qualcomm Code-Excited Linear Prediction), variable bit rate vocoders, are used for mobile phones in 3G environment. For the real-time implementation of a vocoder, the reduction of sound quality is a typical problem. To improve the sound quality, that is important to know the size and shape of noise. In the existing sound quality improvement method, the voice activated is detected or used, or statistical methods are used by the large mount of data. However, there is a disadvantage in that no noise can be detected, when there is a continuous signal or when a change in noise is large.This paper focused on finding a better way to decrease the reduction of sound quality in lower bit transmission environments. Based on simulation results, this study proposed a preprocessor application that estimates the SNR (Signal to Noise Ratio) using the spectral SNR estimation method. The SNR estimation method adopted the IMBE (Improved Multi-Band Excitation) instead of using the SNR, which is a continuous speech signal. Finally, this application improves the quality of the vocoder by enhancing sound quality adaptively.

• 운동영양학회지
• /
• 제17권2호
• /
• pp.25-34
• /
• 2013

This study was conducted to investigate the relationship between ACE gene polymorphism and muscle damage parameters after eccentric exercise. 80 collegiate males were instructed to take an eccentric exercise with the elbow flexor muscle through the modified preacher curl machine for 2 sets of 25 cycles (total 50 cycles). The maximal isometric strength, muscle soreness, creatine kinase (CK), and myoglobin (Mb) were measured before exercise, and 0, 24, 48, 72, and 96 hrs after exercise. The result showed that after the eccentric exercise, the maximal isometric strength significantly decreased by more than 50% (p < 0.001) and the muscle soreness, CK, and Mb significantly increased compared to those before the exercise (p < 0.001). The ACE gene polymorphism of the subjects was classified using real-time polymerase chain reaction (real-time PCR). The result showed that it consisted of 38 cases of type II (46.4%), 33 cases of type ID (43.4%), and 9 cases of type DD (10.2%). The Hardy-Weinberg equilibrium for ACE gene polymorphism was shown to have p = 0.653, which showed that each allele was evenly distributed. Although significant differences in the changes in the maximal isometric strength, muscle soreness, CK, and Mb were found according to time course (p < 0.001), no significant differences in the changes in the maximal isometric strength, muscle soreness, CK, and Mb were found according to ACE gene polymorphism. Furthermore, no significant difference in the changes in the muscle damage parameters was found according to interaction between ACE gene polymorphism and time course (p > 0.05). In conclusion, the level of the muscle damage parameters changed in the injured muscle after eccentric exercise, but these changes in the muscle damage parameters were not affected by ACE gene polymorphism. The result of this study indicates that ACE gene is not a candidate gene that explains muscle damage.

• 한국컴퓨터정보학회논문지
• /
• 제26권11호
• /
• pp.139-147
• /
• 2021

도서관 계에서 기계 가독 목록 형식(MARC)에 대한 대안으로 주목받고 있는 BIBFRAME은 기존 데이터와의 호환성을 유지하면서 오픈 웹 환경에서 새로운 서지기술 데이터모델을 제시한다. MARC의 한국형 데이터 모델인 KORMARC 레코드의 BIBFRAME 변환을 위해, 서울시 노원구립도서관의 최신 서지 데이터 5,000개를 분석하여 25개의 핵심 필드를 추출한다. 핵심 필드들을 MARC 21의 호환성 여부에 따라 세 가지 유형으로 분류하고, 각 유형별 변환 기법을 정의한다. 또한 오픈소스 기반의 변환기를 구현하여 KORMARC to BIBFRAME 변환 작업을 수행한다. 본 연구는 KORMARC to BIBFRAME 변환에 대한 기초연구로써, 실제 사용되는 최신 KORMARC 정보를 분석하여 변환 규칙을 정의하고, BIBFRAME 변환을 시도했다는 점에 의의가 있다.

• 수산해양기술연구
• /
• 제59권1호
• /
• pp.28-34
• /
• 2023

In the anchovy boat seine fishing boat, it is necessary to select other aquatic organisms other than live anchovies, which are the target species of catch. By making a rotating roller sorter using hydraulic pressure, the anchovy sorting amount was compared and the sorting accuracy of the rotary roller sorter, and the discharge speed of butter fish and jerry fish according to the number of roller revolutions were analyzed. The rotating roller sorter increases the weight of the sorted raw anchovy by 54%, 74% and 91.5% compared to the round bar fixed type, so it can reduce the required time by an average of 73.2%. As a result of converting the sorting accuracy to the weight of pure anchovies excluding the catch weight, the round bar fixed type was 89%; however, the average of the rotating roller sorter was 97.7%. Thus, the sorting accuracy of the rotary roller sorter was further improved by about 8.7%. The roller speed moved 7% at 300 rpm, 7.5% at 600 rpm, and 16% at 900 rpm, so butter fish were discharged overboard 10% faster than jelly fish on average. In addition, the average feed speed of butter fish and jelly fish is 1,400 mm/s when the roller rotation speed is 300 rpm, 1,480 mm/s at 600 rpm, and 1,850 mm/s at 900 rpm. A Φ58 mm roller rotates once it moved about 1.23 mm. In the future, a follow-up study of quantitative evaluation is needed targeting more non-target fish species of anchovy boat seine.

• International Journal of Computer Science & Network Security
• /
• 제24권2호
• /
• pp.101-112
• /
• 2024

Telugu language is considered as fourth most used language in India especially in the regions of Andhra Pradesh, Telangana, Karnataka etc. In international recognized countries also, Telugu is widely growing spoken language. This language comprises of different dependent and independent vowels, consonants and digits. In this aspect, the enhancement of Telugu Handwritten Character Recognition (HCR) has not been propagated. HCR is a neural network technique of converting a documented image to edited text one which can be used for many other applications. This reduces time and effort without starting over from the beginning every time. In this work, a Unicode based Handwritten Character Recognition(U-HCR) is developed for translating the handwritten Telugu characters into English language. With the use of Centre of Gravity (CG) in our model we can easily divide a compound character into individual character with the help of Unicode values. For training this model, we have used both online and offline Telugu character datasets. To extract the features in the scanned image we used convolutional neural network along with Machine Learning classifiers like Random Forest and Support Vector Machine. Stochastic Gradient Descent (SGD), Root Mean Square Propagation (RMS-P) and Adaptative Moment Estimation (ADAM)optimizers are used in this work to enhance the performance of U-HCR and to reduce the loss function value. This loss value reduction can be possible with optimizers by using CNN. In both online and offline datasets, proposed model showed promising results by maintaining the accuracies with 90.28% for SGD, 96.97% for RMS-P and 93.57% for ADAM respectively.

• 문화기술의 융합
• /
• 제10권1호
• /
• pp.471-476
• /
• 2024

다양한 스마트 기기 및 컴퓨팅 디바이스의 보급에 따라 빅데이터 생성이 광범위하게 일어나고 있다. 기계학습은 데이터의 패턴을 학습하여 추론을 수행하는 알고리즘이다. 다양한 기계학습 알고리즘 중에서 주목을 받는 알고리즘은 신경망 기반의 딥러닝 학습이다. 딥러닝은 다양한 응용이 발표되면서 빠른 성능 향상을 달성하고 있다. 최근 딥러닝 알고리즘 중에서 그래프 구조를 활용하여 데이터를 분석하려는 시도가 증가하고 있다. 본 연구에서는 그래프 구조를 활용하여 딥러닝 네트워크에 전달하기 위한 그래프 생성 방법을 제시한다. 본 논문은 그래프 생성 과정에서 노드의 속성과 간선의 가중치를 일반화하고 행렬화 과정을 제시하여 딥러닝 입력에 필요한 구조로 전환하는 방법을 제시한다. 그래프 생성 과정에서 속성과 가중치 정보를 보전할 수 있는 선형변환 매트릭스 적용 방법을 제시한다. 마지막으로 일반 그래프의 딥러닝 입력 구조를 제시하고 성능 분석을 위한 접근법을 제시한다.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제12권2호
• /
• pp.99-110
• /
• 2023

본 연구는 현 보안 관제 시스템이 직면한 실시간 트래픽 탐지 문제를 해결하기 위해 사이버 위협 프레임워크인 마이터 어택과 머신러닝을 이용하여 유해 네트워크 트래픽을 분류하는 방안을 제안하였다. 마이터 어택 프레임워크에 네트워크 트래픽 데이터셋인 UNSW-NB15를 적용하여 라벨을 변환 후 희소 클래스 처리를 통해 최종 데이터셋을 생성하였다. 생성된 최종 데이터셋을 사용하여 부스팅 기반의 앙상블 모델을 학습시킨 후 이러한 앙상블 모델들이 다양한 성능 측정 지표로 어떻게 네트워크 트래픽을 분류하는지 평가하였다. 그 결과 F-1 스코어를 기준으로 평가하였을 때 희소 클래스 미처리한 XGBoost가 멀티 클래스 트래픽 환경에서 가장 우수함을 보였다. 학습하기 어려운 소수의 공격클래스까지 포함하여 마이터 어택라벨 변환 및 오버샘플링처리를 통한 머신러닝은 기존 연구 대비 차별점을 가지고 있으나, 기존 데이터셋과 마이터 어택 라벨 간의 변환 시 완벽하게 일치할 수 없는 점과 지나친 희소 클래스 존재로 인한 한계가 있음을 인지하였다. 그럼에도 불구하고 B-SMOTE를 적용한 Catboost는 0.9526의 분류 정확도를 달성하였고 이는 정상/비정상 네트워크 트래픽을 자동으로 탐지할 수 있을 것으로 보인다.