• International conference on construction engineering and project management
• /
• 2024.07a
• /
• pp.327-334
• /
• 2024

Bridge inspection is crucial for infrastructure maintenance. Current inspections based on computer vision primarily focus on identifying simple defects such as cracks or corrosion. These detection results can serve merely as preliminary references for bridge inspection reports. To generate detailed reports, on-site engineers must still present the structural conditions through lengthy textual descriptions. This process is time-consuming, costly, and prone to human error. To bridge this gap, we propose a deep learning-based framework to generate detailed and accurate textual descriptions, laying the foundation for automating bridge inspection reports. This framework is built around an encoder-decoder architecture, utilizing Convolutional Neural Networks (CNN) for encoding image features and Gated Recurrent Units (GRU) as the decoder, combined with a dynamically adaptive attention mechanism. The experimental results demonstrate this approach's effectiveness, proving that the introduction of the attention mechanism contributes to improved generation results. Moreover, it is worth noting that, through comparative experiments on image restoration, we found that the model requires further improvement in terms of explainability. In summary, this study demonstrates the potential and practical application of image captioning techniques for bridge defect detection, and future research can further explore the integration of domain knowledge with artificial intelligence (AI).

• Journal of Advanced Marine Engineering and Technology
• /
• v.30 no.4
• /
• pp.520-526
• /
• 2006

Wave Energy is a derivative of the solar energy input to the earth, which is accumulated on open water surfaces by the action of the winds Waves are disturbances in the water surface. This paper is interested primarily in progressive waves, which carry energy from one place to another Waves are irregular in size and frequency. Moreover the surface of the sea is one of the most hostile environments for engineering structures and materials. The idea of harnessing the tremendous power of the ocean's waves is not new. Hundreds of wave energy conversion techniques have been suggested over the last two centuries. Although many WECS (Wave Energy Conversion Systems) have been invented, only a few systems have been tested and evaluated. This paper describes the characteristic of WES (Wave Energy System) in terms of, devices, resource and potential, etc.. Finally, this paper provides a summary of general and specific conclusions and recommendations concerning WECS potential in Korea.

• Genomics & Informatics
• /
• v.14 no.4
• /
• pp.211-215
• /
• 2016

The alteration of alternative splicing patterns has an effect on the quantification of functional proteins, leading to phenotype variation. The splicing quantitative trait locus (sQTL) is one of the main genetic elements affecting splicing patterns. Here, we report the results of genome-wide sQTLs across 141 strains of Arabidopsis thaliana with publicly available next generation sequencing datasets. As a result, we found 1,694 candidate sQTLs in Arabidopsis thaliana at a false discovery rate of 0.01. Furthermore, among the candidate sQTLs, we found 25 sQTLs that overlapped with the list of previously examined trait-associated single nucleotide polymorphisms (SNPs). In summary, this sQTL analysis provides new insight into genetic elements affecting alternative splicing patterns in Arabidopsis thaliana and the mechanism of previously reported trait-associated SNPs.

• Annual Conference of KIPS
• /
• 2017.11a
• /
• pp.546-548
• /
• 2017

오픈소스인 Scouter를 바탕으로 제작한, 기업의 WAS와 OS의 안정적인 서비스를 위한 실시간 성능 모니터링 시스템이다. 기존 Scouter 사용에서 불편함이 제기되었던 알림의 빈도 및 간격 조절 기능을 추가하였다. 더불어 요약 및 보고의 편의성을 돕기 위해 그래프를 포함한 PDF 형식의 성능 요약 보고서를 주기별로 생성하여 Email로 전송해주는 자동 성능 요약 보고서 생성 모듈을 새롭게 개발하였다. 본 논문에서는 이와 같은 성능개선 및 신기술 개발에 관련한 활동에 대해 논하겠다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.04b
• /
• pp.577-579
• /
• 2002

최근 디지털 동명상의 사용이 증가하면서 자동으로 구조를 분석하는 기술이 필요하게 되었다. 특히 뉴스 동영상은 다른 동영상에 비해 그 구조가 비교적 정명화 되어 있다는 특징 때문에 많은 연구에서 이용되어졌다. 뉴스 동영상을 이용하는 이전의 연구에서 문제점으로 제시될 수 있는 사항은 서로 다른 자료 구조와 시스템 구조로 인 딴 호환성의 부족을 들 수 있다. 본 논문에서 는 이와 같은 호환성 부족을 해결하기 위해 멀티미디어 데이터를 기술하는 표준인 MPEG-7을 기반으로 한 뉴스 동영상 스키마를 제안하였다 특히 전제 뉴스를 보지 않고 효율적으로 뉴스 동영상을 이해할 수 있도록 요약하는 방법을 제시하였으며 MPEG-7의 HDS를 이용하여 기술하였다. 본 논문에서 제안한 방법은 디지털 비디오 라이브러리와 같은 응용 분야에서 유용하게 이용될 수 있을 것이다.

• Journal of Microbiology and Biotechnology
• /
• v.25 no.2
• /
• pp.145-151
• /
• 2015

Graphene is a next-generation biomaterial with increasing biomedical applicability. As a new class of one-atom-thick nanosheets, it is a true two-dimensional honeycomb network nanomaterial that attracts interest in various scientific fields and is rapidly becoming the most widely studied carbon-based material. Since its discovery in 2004, its unique optical, mechanical, electronic, thermal, and magnetic properties are the basis of exploration of the potential applicability of graphene. Graphene materials, such as graphene oxide and its reduced form, are studied extensively in the biotechnology arena owing to their multivalent functionalization and efficient surface loading with various biomolecules. This review provides a brief summary of the recent progress in graphene and graphene oxide biological research together with current findings to spark novel applications in biomedicine. Graphene-based applications are progressively developing; hence, the opportunities and challenges of this rapidly growing field are discussed together with the versatility of these multifaceted materials.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.530-531
• /
• 2018

문서 자동 요약은 주요 단어 또는 문장을 추출하거나 문장을 생성하는 방식으로 요약한다. 최근 연구에서는 대량의 문서를 딥러닝하여 요약문 자체를 생성하는 방식으로 발전하고 있다. 추출 요약이나 생성 요약 모두 핵심 단어를 인식하는 것이 매우 중요하다. 학습할 때 각 단어가 문장에서 출현한 패턴으로부터 의미를 인식하고 단어를 선별하여 요약한다. 결국 기계학습에서는 학습 문서에 출현한 어휘만으로 요약을 한다. 따라서 학습 문서에 출현하지 않았던 어휘가 포함된 새로운 문서의 요약에서 기존 모델이 잘 작동하기 어려운 문제가 있다. 본 논문에서는 학습단계에서 출현하지 않은 단어까지도 중요성을 인식하고 요약문을 생성할 수 있는 신경망 모델을 제안하였다.

• Current Optics and Photonics
• /
• v.1 no.6
• /
• pp.631-636
• /
• 2017

We firstly develop a straightforward method to generate full $Poincar{\acute{e}}$ beams with any polarization geometry over an arbitrary order $Poincar{\acute{e}}$ sphere. We implement this by coaxial superposition of two orthogonal circular polarized beams with alternative topological charges with the help of a Mach-Zehnder interferometer. Secondly we find the existence of singularity points. And the inner relationship between their characteristics and the order of $Poincar{\acute{e}}$ spheres is also studied. In summary, this work provides a convenient and effective way to generate vector beams and to control their polarization states.

• Proceedings of the Korea Institute of Fire Science and Engineering Conference
• /
• 1997.11a
• /
• pp.361-368
• /
• 1997

Summary of experiments for the investigation of a fire which caused an upward fire spread for over 12 floors through balconies in a high-rise apartment complex is reported. The experiments include indoor tests to obtain fire properties of vertical PMMA fences and outdoor ones with a full scale model of the balcony. The test results suggest significance of the increase of total flame height by the merging of flames and a cooperative effect of the burning of the PMMA fence and combustibles on the balconies for the generation of a tall flame enough to cause ignition on the upper floors.

• Korean Journal of Radiology
• /
• v.22 no.5
• /
• pp.677-687
• /
• 2021

Microvascular ultrasound (US) techniques are advanced Doppler techniques that provide high sensitivity and spatial resolution for detailed visualization of low-flow vessels. Microvascular US imaging can be applied to breast lesion evaluation with or without US contrast agents. Microvascular US imaging without a contrast agent uses a sophisticated wall filtering system to selectively obtain low-flow Doppler signals from overlapped artifacts. Microvascular US imaging with second-generation contrast agents amplifies flow signals and makes them last longer, which facilitates hemodynamic evaluation of breast lesions. In this review article, we will introduce various microvascular US techniques, explain their clinical applications in breast cancer diagnosis and radiologic-histopathologic correlation, and provide a summary of a recent radiogenomic study using microvascular US.