• 대한의용생체공학회:의공학회지
• /
• 제42권4호
• /
• pp.167-174
• /
• 2021

The purpose of this study is to classify TIFF images, PNG images, and JPEG images using deep learning, and to compare the accuracy by verifying the classification performance. The TIFF, PNG, and JPEG images converted from chest X-ray DICOM images were applied to five deep neural network models performed in image recognition and classification to compare classification performance. The data consisted of a total of 4,000 X-ray images, which were converted from DICOM images into 16-bit TIFF images and 8-bit PNG and JPEG images. The learning models are CNN models - VGG16, ResNet50, InceptionV3, DenseNet121, and EfficientNetB0. The accuracy of the five convolutional neural network models of TIFF images is 99.86%, 99.86%, 99.99%, 100%, and 99.89%. The accuracy of PNG images is 99.88%, 100%, 99.97%, 99.87%, and 100%. The accuracy of JPEG images is 100%, 100%, 99.96%, 99.89%, and 100%. Validation of classification performance using test data showed 100% in accuracy, precision, recall and F1 score. Our classification results show that when DICOM images are converted to TIFF, PNG, and JPEG images and learned through preprocessing, the learning works well in all formats. In medical imaging research using deep learning, the classification performance is not affected by converting DICOM images into any format.

• 인터넷정보학회논문지
• /
• 제20권4호
• /
• pp.91-102
• /
• 2019

자율주행 시스템에서, 카메라에 포착된 영상을 통하여 보행자를 분류하는 기능은 보행자 안전을 위하여 매우 중요하다. 기존에는 HOG(Histogram of Oriented Gradients)나 SIFT(Scale-Invariant Feature Transform) 등으로 보행자의 특징을 추출한 후 SVM(Support Vector Machine)으로 분류하는 기술을 사용했었으나, 보행자 특징을 위와 같이 수동(handcrafted)으로 추출하는 것은 많은 한계점을 가지고 있다. 따라서 본 논문에서는 CNN(Convolutional Neural Network)의 깊은 특징(deep features)과 전이학습(transfer learning)을 사용하여 보행자를 안정적이고 효과적으로 분류하는 방법을 제시한다. 본 논문은 2가지 대표적인 전이학습 기법인 고정특징추출(fixed feature extractor) 기법과 미세조정(fine-tuning) 기법을 모두 사용하여 실험하였고, 특히 미세조정 기법에서는 3가지 다른 크기로 레이어를 전이구간과 비전이구간으로 구분한 후, 비전이구간에 속한 레이어들에 대해서만 가중치를 조정하는 설정(M-Fine: Modified Fine-tuning)을 새롭게 추가하였다. 5가지 CNN모델(VGGNet, DenseNet, Inception V3, Xception, MobileNet)과 INRIA Person데이터 세트로 실험한 결과, HOG나 SIFT 같은 수동적인 특징보다 CNN의 깊은 특징이 더 좋은 성능을 보여주었고, Xception의 정확도(임계치 = 0.5)가 99.61%로 가장 높았다. Xception과 유사한 성능을 내면서도 80% 적은 파라메터를 학습한 MobileNet이 효율성 측면에서는 가장 뛰어났다. 그리고 3가지 전이학습 기법중 미세조정 기법의 성능이 가장 우수하였고, M-Fine 기법의 성능은 미세조정 기법과 대등하거나 조금 낮았지만 고정특징추출 기법보다는 높았다.

• 한국산학기술학회논문지
• /
• 제21권12호
• /
• pp.320-330
• /
• 2020

화재는 높은 비정형성으로 인해 딥러닝 모델을 이용한 영상인식 분야에서도 좋은 성능을 내기가 어려운 대상 중 하나이다. 특히 지하공동구 내 화재는 딥러닝 모델의 학습을 위한 화재 데이터 확보가 어렵고 열약한 영상 조건 및 화재로 오인할 수 있는 객체가 많아 화재 검출이 어렵고 성능이 낮다. 이러한 이유로 본 연구는 딥러닝 기반의 지하공동구 내 화재 탐지 모델을 제안하고, 제안된 모델의 성능을 평가하였다. 기존 합성곱 인공신경망에 GoogleNet의 Inception block과 ResNet의 skip connection을 조합하여 어두운 환경에서 발생되는 화재 탐지를 위한 모델 구조를 제안하였으며, 제안된 모델을 효과적으로 학습시키기 위한 방법도 함께 제시하였다. 제안된 방법의 효과를 평가하기 위해 학습 후 모델을 지하공동구 및 유사환경 조건의 화재 문제와 화재로 오인할 수 있는 객체를 포함한 이미지에 적용해 결과를 분석하였다. 또한 기존 딥러닝 기반 화재 탐지 모델의 정밀도, 검출률 지표와 비교함으로써 모델의 화재 탐지 성능을 정량적으로 평가하였다. 제안된 모델의 결과는 어두운 환경에서 발생되는 화재 문제에 대해 높은 정밀도와 검출률을 나타내었으며, 유사 화재 객체에 대해 낮은 오탐 및 미탐 성능을 가지고 있음을 보여주었다.

• 디지털융복합연구
• /
• 제13권1호
• /
• pp.135-149
• /
• 2015

본 연구의 목적은 샌디에고 바이오 클러스트를 형성하는데 핵심적 역할을 한 CONNECT의 성공요인과 프로그램들을 문헌연구를 통해 분석하고 소개함으로서 우리나라 창업보육센터의 서비스를 향상시키기 위한 방향성을 제시하는데 있다. 현재 CONNECT는 창업보육에 있어 세계적인 벤치마킹사례로 유명하다. 본 논문에서는 국내의 지방자치단체 및 창업보육기관들이 쉽게 벤치마킹할 수 있도록 CONNECT의 성공요인을 분석하고 핵심보육프로그램들을 소개하는데 중점을 두었다. UCSD CONNECT는 기업성장단계별로 다양한 기업의 필요에 따른 지원프로그램을 잘 갖추고 있으며 현재까지 800개가 넘는 하이테크기업들을 육성했다. CONNECT는 하드웨어적 지원방식에 익숙한 우리나라의 창업보육센터가 향후 중점을 두어야 할 다양한 소프트웨어적 지원 프로그램들을 제공해 왔다. 정책책임자의 리더쉽, 네트워크 및 소프트웨어적 기업지원서비스가 UCSD CONNECT의 가장 중요한 성공요인이었다.

• 대한간호학회지
• /
• 제30권5호
• /
• pp.1099-1110
• /
• 2000

This paper aims to examine what nursing discipline has accomplishd to date and projects what could be its preferred future from global perspective. Major contextual factors that influence nursing are examined in light of their significance on the progress of nursing discipline. These include evolution of society, and trends in higher education and health care market. The perspective of world health is gained from WHO, an organization recognized for its mission for the health of people worldwide. As the future builds on the present that, in turn, builds on the past, major milestones of nursing discipline, particularly that of education system from the inception of nursing to present is highlighted. The importance of research to advance science and improve peoples health are presented along with a call for nursing research to be responsive to societal needs. The preferred future for nursing discipline is presented integrating the trends of society, higher education, and health care environment. Doctoral education that is the hallmark of nursing scholarship is further elaborated in terms of its mission, needs, and quality attainment. Data from the International Network of Doctoral Education in Nursing are presented along with information about current attempts in developing quality criteria and indicators for doctoral education in nursing worldwide. Majority of information in this paper comes from the United States, unless specified otherwise.

• The Journal of Asian Finance, Economics and Business
• /
• 제7권8호
• /
• pp.695-704
• /
• 2020

Bitcoin and other prominent cryptocurrencies have gained much attention since the last several years. Globally known as digital coin and virtual currency, this cryptocurrency is gained and traded within the blockchain system. The blockchain technology adopted in using the cryptocurrency has raised the eyebrows within the banking sector, government, stakeholders and individual investors. The rise of the cryptocurrency within this decade since the inception of Bitcoin in 2009 has taken the market by storm. Cryptocurrency is anticipated as the future currency that might replace the current paper currency worldwide. Even though the interest has caught the attention of users, many are not aware of its opportunities, drawbacks and challenges for the future. Researches on cryptocurrencies are still lacking and still at its infancy stage. In providing substantial guide and view to the academic field and users, this paper will discuss the opportunities in the cryptocurrency such as the security of its technology, low transaction cost and high investment return. The originality of this paper is on the discussion within law and regulation, high energy consumption, possibility of crash and bubble, and attacks on network. The future undertakings of cryptocurrency and its application will be systematically reviewed in this paper.

• Journal of Electrical Engineering and Technology
• /
• 제13권5호
• /
• pp.1807-1813
• /
• 2018

In this paper, the fault line selection and location problems of single line-to-ground (SLG) fault in distribution network are addressed. Firstly, the adaptive filtering property for empirical mode decomposition is formulated. Then in view of the different characteristics showed by the intrinsic mode functions(IMF) under different fault inception angles obtained by empirical mode decomposition, the sign of peak value about the low-frequency IMF and the capacitance transient energy is chosen as the fault line selection criteria according to the different proportion occupied by the low-frequency components. Finally, the fault location is determined based upon the comparison result with adjacent fault passage indicators' (FPI) waveform on the strength of the interaction between the distribution terminal unit(DTU) and the FPI. Moreover, the logic nodes regarding to fault line selection and location are newly expanded according to IEC61850, which also provides reference to acquaint the DTU or FPI's function and monitoring. The simulation results validate the effectiveness of the proposed fault line selection and location methods.

• 한국전기전자재료학회논문지
• /
• 제37권4호
• /
• pp.347-357
• /
• 2024

Oxide semiconductor gas sensors are widely used for detecting toxic, explosive, and flammable gases due to their simple structure, cost-effectiveness, and potential integration into compact devices. However, their reliable gas detection is hindered by a longstanding issue known as humidity dependence, wherein the sensor resistance and gas response change significantly in the presence of moisture. This problem has persisted since the inception of oxide semiconductor gas sensors in the 1960s. This paper explores the root causes of humidity dependence in oxide semiconductor gas sensors and presents strategies to address this challenge. Mitigation strategies include functionalizing the gas-sensing material with noble metal/transition metal oxides and rare-earth/rare-earth oxides, as well as implementing a moisture barrier layer to prevent moisture diffusion into the gas-sensing film. Developing oxide semiconductor gas sensors immune to humidity dependence is expected to yield substantial socioeconomic benefits by enabling medical diagnosis, food quality assessment, environmental monitoring, and sensor network establishment.

• 한국멀티미디어학회논문지
• /
• 제22권5호
• /
• pp.613-625
• /
• 2019

The rise of the Internet of Things (IoT) devices have greatly influenced many industries and one of them is healthcare where wearable devices started to track all your daily activities for better health monitoring accuracy and even down to tracking daily food intake in some cases. With the amounts of data that are being tracked and shared between from these devices, questions were raised on how to uphold user's data privacy when data is shared between these IoT devices and third party. With the blockchain platforms started to mature since its inception, the technology can be implemented according to a variety of use case scenarios. In this paper, we present a system architecture based on the healthcare system and IoT network by leveraging on multiple blockchain networks as the medium in between that should enable users to have direct authority on data accessibility of their shared data. We provide proof of concept implementation and highlight the results from our testing to show how the efficiency and scalability of the healthcare system improved without having a significant impact on the performance of the Electronic Medical Record (EMR) that mostly affected by the previous solution since these solutions directly connected to a public blockchain network and which resulted in significant delays and high cost of operation when a large amount of data or complicated functions are involved.

• Journal of Acupuncture Research
• /
• 제40권3호
• /
• pp.238-244
• /
• 2023

Background: Occipital neuralgia (ON) is an established risk factor for headaches in the posterior cervical region. Several conservative treatments by nerve decompression and pain relief are available for ON, but these treatments have limitations. Acupuncture treatment, which is known to demonstrate analgesic effects, involves various stimulation methods, and several studies have reported their clinical benefit. No recent systematic review (SR) has compared each acupuncture type for ON treatment. Thus, this SR aims to investigate the clinical effectiveness of each acupuncture type for treating ON. Methods: We will identify relevant studies using electronic databases, including EMBASE, MEDLINE, Cochrane Library, China National Knowledge Infrastructure (CNKI), Korean Studies Information Service System (KISS), Korean Medical Database, KoreaMed, and National Digital Science Library (NDSL) from the inception until August 2023. The primary outcome will include the numerical change of pain symptoms (visual analog scale and numerical rating scale) and effective rate. Safety and secondary outcomes will include adverse events and quality of life. We will compare the conservative treatment with the acupuncture treatment using network meta-analysis. The Cochrane Collaboration "risk of bias" tools will be used to assess the quality of included trials. The Grades of Recommendation, Assessment, Development, and Evaluation will be used to examine the evidence level. Conclusion: This study will provide clinical evidence of several acupuncture types for ON and help clinicians decide on the best.