Fender segmentation in unmanned aerial vehicle images based on densely connected receptive field block

Yu Byeongjun,전해민,방현태,Yi Sang Soo,Min Jiyoung 대한조선학회 2022 International Journal of Naval Architecture and Oc Vol.14 No.1

Periodic evaluation of port structures including outer walls and berthing structures with fenders is paramount to ensure their safety. However, most fenders are inaccessible on land and inspectors utilize floating boats for inspections. Therefore, this study aimed to develop a fender segmentation system incorporating a vision sensor with deep learning segmentation approach. The semantic segmentation model in an encoderedecoder framework was densely connected through a receptive field block convolution module each with two different dilation rates and classified various types and sizes of fenders efficiently. The images of various types of fenders were assembled and augmented to train the proposed network, and its performance was compared to conventional segmentation models. The trained network was finally applied to unmanned aerial vehicle images and the results showed that fenders were successfully segmented even in the case of the images that include changes in shapes or colors.

정적 분석에서의 거짓 호출 사이클의 존재

유병준(Byeongjun Yu),오학주(Hakjoo Oh),이광근(Kwangkeun Yi) 한국정보과학회 2012 한국정보과학회 학술발표논문집 Vol.39 No.2A

암시적 계산 언어 실행기의 구현

유병준(Byeongjun Yu),이광근(Kwangkeun Yi) 한국정보과학회 2011 한국정보과학회 학술발표논문집 Vol.38 No.2A

Haskell의 타입 클래스(Type Class) 등과 같이 타입 정보를 통해 프로그램 코드를 자동으로 넘겨주는 프로그래밍 기술이 유용하게 쓰이고 있다. 타입 추론이 코드를 주면 자동으로 그 코드의 타입을 유추하는 것처럼 암시적 계산 언어는 주어진 코드의 타입을 보고 필요한 코드가 무엇인지 유추해 낼 수 있다. 본 논문에서는 위와 같은 암시적인 계산 언어들의 주요한 성질을 뽑아 만든 언어의 실행기를 만들었다. 또 실행기를 구현함에 있어서 일반적인 실행기에서 거의 쓰이지 않는 동일화 알고리즘(Unification Algorithm)을 사용했다.

High-Performance n-Type PbSe-Cu₂Se Thermoelectrics through Conduction Band Engineering and Phonon Softening

Zhou, Chongjian,Yu, Yuan,Lee, Yong Kyu,Cojocaru-Miré,din, Oana,Yoo, Byeongjun,Cho, Sung-Pyo,Im, Jino,Wuttig, Matthias,Hyeon, Taeghwan,Chung, In American Chemical Society 2018 JOURNAL OF THE AMERICAN CHEMICAL SOCIETY - Vol.140 No.45

<P>From a structural and economic perspective, tellurium-free PbSe can be an attractive alternative to its more expensive isostructural analogue of PbTe for intermediate temperature power generation. Here we report that PbSe<SUB>0.998</SUB>Br<SUB>0.002</SUB>-2%Cu<SUB>2</SUB>Se exhibits record high peak <I>ZT</I> 1.8 at 723 K and average <I>ZT</I> 1.1 between 300 and 823 K to date for all previously reported n- and p-type PbSe-based materials as well as tellurium-free n-type polycrystalline materials. These even rival the highest reported values for n-type PbTe-based materials. Cu<SUB>2</SUB>Se doping not only enhance charge transport properties but also depress thermal conductivity of n-type PbSe. It flattens the edge of the conduction band of PbSe, increases the effective mass of charge carriers, and enlarges the energy band gap, which collectively improve the Seebeck coefficient markedly. This is the first example of manipulating the electronic conduction band to enhance the thermoelectric properties of n-type PbSe. Concurrently, Cu<SUB>2</SUB>Se increases the carrier concentration with nearly no loss in carrier mobility, even increasing the electrical conductivity above ∼423 K. The resulting power factor is ultrahigh, reaching ∼21-26 μW cm<SUP>-1</SUP> K<SUP>-2</SUP> over a wide range of temperature from ∼423 to 723 K. Cu<SUB>2</SUB>Se doping substantially reduces the lattice thermal conductivity to ∼0.4 W m<SUP>-1</SUP> K<SUP>-1</SUP> at 773 K, approaching its theoretical amorphous limit. According to first-principles calculations, the achieved ultralow value can be attributed to remarkable acoustic phonon softening at the low-frequency region.</P> [FIG OMISSION]</BR>

Assembly performance evaluation method for prefabricated steel structures using deep learning and k-nearest neighbors

Haemin Jeon,Hyuntae Bang,Byeongjun Yu 국제구조공학회 2023 Smart Structures and Systems, An International Jou Vol.32 No.2

This study proposes an automated assembly performance evaluation method for prefabricated steel structures (PSSs) using machine learning methods. Assembly component images were segmented using a modified version of the receptive field pyramid. By factorizing channel modulation and the receptive field exploration layers of the convolution pyramid, highly accurate segmentation results were obtained. After completing segmentation, the positions of the bolt holes were calculated using various image processing techniques, such as fuzzy-based edge detection, Hough's line detection, and image perspective transformation. By calculating the distance ratio between bolt holes, the assembly performance of the PSS was estimated using the k-nearest neighbors (kNN) algorithm. The effectiveness of the proposed framework was validated using a 3D PSS printing model and a field test. The results indicated that this approach could recognize assembly components with an intersection over union (IoU) of 95% and evaluate assembly performance with an error of less than 5%.

Prevention of Bacterial Colonization on Catheters by a One-Step Coating Process Involving an Antibiofouling Polymer in Water

Keum, Hyeongseop,Kim, Jin Yong,Yu, Byeongjun,Yu, Seung Jung,Kim, Jinjoo,Jeon, Hyungsu,Lee, Dong Yun,Im, Sung Gap,Jon, Sangyong American Chemical Society 2017 ACS APPLIED MATERIALS & INTERFACES Vol.9 No.23

<P>As reports of multidrug resistant pathogens have increased, patients with implanted medical catheters increasingly need alternative solutions to antibiotic treatments. As most catheter-related infections are directly associated with biofilm formation on the catheter surface, which, once formed, is difficult to eliminate, a promising approach to biofilm prevention involves inhibiting the initial adhesion of bacteria to the surface. In this study, we report an amphiphilic, antifouling polymer, poly(DMA-mPEGMA-AA) that can facilely coat the surfaces of commercially available catheter materials in water and prevent bacterial adhesion to and subsequent colonization of the surface, giving rise to an antibiofilm surface. The antifouling coating layer was formed simply by dipping a model substrate (polystyrene, PET, PDMS, or silicon-based urinary catheter) in water containing poly(DMA-mPEGMA-AA), followed by characterization by X-ray photoelectron spectroscopy (XPS). The antibacterial adhesion properties of the polymer-coated surface were assessed for Staphylococcus aureus (S. aureus) and Escherichia coli (E. coli) growth under static (incubation in the presence of a bacterial suspension) and dynamic (bacteria suspended in a solution under flow) conditions. Regardless of the conditions, the polymer-coated surface displayed significantly reduced attachment of the bacteria (antiadhesion effect > similar to 8-fold) compared to the bare noncoated substrates. Treatment of the implanted catheters with S. aureus in vivo further confirmed that the polymer-coated silicon urinary catheters could significantly reduce bacterial adhesion and biofilm formation in a bacterial infection animal model. Furthermore, the polymer-coated catheters did not induce hemolysis and were resistant to the adhesion of blood-circulating cells, indicative of high biocompatibility. Collectively, the present amphiphilic antifouling polymer is potentially useful as a coating platform that renders existing medical devices resistant to biofilm formation.</P>

Ultrasensitive Electrochemical Detection of miRNA-21 Using a Zinc Finger Protein Specific to DNA–RNA Hybrids

Fang, Chiew San,Kim, Kwang-sun,Yu, Byeongjun,Jon, Sangyong,Kim, Moon-Soo,Yang, Haesik American Chemical Society 2017 ANALYTICAL CHEMISTRY - Vol.89 No.3

<P>Both high sensitivity and high specificity are crucial for detection of miRNAs that have emerged as important clinical biomarkers. Just Another Zinc finger proteins (JAZ, ZNF346) bind preferably (but nonsequence-specifically) to DNA RNA hybrids over single stranded RNAs, single stranded DNAs, and double stranded DNAs. We present an ultrasensitive and highly specific electrochemical method for miRNA-21 detection based on the selective binding of JAZ to the DNA RNA hybrid formed between a DNA capture probe and a target miRNA-21. This enables us to use chemically stable DNA as a capture probe instead of RNA as well as to apply a standard sandwich type assay format to miRNA detection. High signal amplification is obtained by (i) enzymatic amplification by alkaline phosphatase (ALP) coupled with (ii) electrochemical chemical chemical (ECC) redox cycling involving an ALP product (hydroquinone). Low nonspecific adsorption of ALP conjugated JAZ is obtained using a polymeric self-assembled-monolayer-modified and casein treated indium tin oxide electrode. The detection method can discriminate between target miRNA-21 and nontarget nucleic acids (DNA DNA hybrid, single stranded DNA, miRNA-125b, miRNA-155, single base mismatched miRNA, and three base mismatched miRNA). The detection limits for miRNA-21 in buffer and 10 fold diluted serum are approximately 2 and 30 fM, respectively, indicating that the detection method is ultrasensitive. This detection method can be readily extended to multiplex detection of miRNAs with only one ALP conjugated JAZ probe due to its nonsequence-specific binding character. We also believe that the method could offer a promising solution for point of care testing of miRNAs in body fluids.</P>

k-근접 이웃 및 비전센서를 활용한 프리팹 강구조물 조립 성능 평가 기술

방현태,유병준,전해민,Bang, Hyuntae,Yu, Byeongjun,Jeon, Haemin 한국전산구조공학회 2022 한국전산구조공학회논문집 Vol.35 No.5

본 논문에서는 프리팹 구조물의 품질관리를 위한 딥러닝 및 비전센서 기반의 조립 성능 평가 모델을 개발하였다. 조립부 검출을 위해 인코더-디코더 형식의 네트워크와 수용 영역 블록 합성곱 모듈을 적용한 딥러닝 모델을 사용하였다. 검출된 조립부 영역 내의 볼트홀을 검출하고, 볼트홀의 위치 값을 산정하여 k-근접 이웃 기반 모델을 사용하여 조립 품질을 평가하였다. 제안된 기법의 성능을 검증하기 위해 조립부 모형을 3D 프린팅을 이용하여 제작하여 조립부 검출 및 조립 성능 예측 모델의 성능을 검증하였다. 성능 검증 결과 높은 정밀도로 조립부를 검출하였으며, 검출된 조립부내의 볼트홀의 위치를 바탕으로 프리팹 구조물의 조립 성능을 5% 이하의 판별 오차로 평가할 수 있음을 확인하였다. In this study, we developed a deep learning and vision sensor-based assembly performance evaluation method isfor prefabricated steel structures. The assembly parts were segmented using a modified version of the receptive field block convolution module inspired by the eccentric function of the human visual system. The quality of the assembly was evaluated by detecting the bolt holes in the segmented assembly part and calculating the bolt hole positions. To validate the performance of the evaluation, models of standard and defective assembly parts were produced using a 3D printer. The assembly part segmentation network was trained based on the 3D model images captured from a vision sensor. The sbolt hole positions in the segmented assembly image were calculated using image processing techniques, and the assembly performance evaluation using the k-nearest neighbor algorithm was verified. The experimental results show that the assembly parts were segmented with high precision, and the assembly performance based on the positions of the bolt holes in the detected assembly part was evaluated with a classification error of less than 5%.

전기차 무선 충전을 위한 주차위치 탐색에 관한 연구

김준영(Junyeong Kim),이동규(Dongkyu Lee),유병준(Byeongjun Yu),기석철(Seok-Cheol Kee) 한국자동차공학회 2020 한국 자동차공학회논문집 Vol.28 No.7

Enhanced Electron Transfer Mediated by Conjugated Polyelectrolyte and Its Application to Washing-Free DNA Detection

Park, Seonhwa,Jeong, Ji-Eun,Le, Van Sang,Seo, Jeongwook,Yu, Byeongjun,Kim, Da-Young,Kwon, Se-Hun,Jon, Sangyong,Woo, Han Young,Yang, Haesik American Chemical Society 2018 JOURNAL OF THE AMERICAN CHEMICAL SOCIETY - Vol.140 No.7

<P>Direct electron transfer between a redox label and an electrode requires a short working distance (<1–2 nm), and in general an affinity biosensor based on direct electron transfer requires a finely smoothed Au electrode to support efficient target binding. Here we report that direct electron transfer over a longer working distance is possible between (i) an anionic π-conjugated polyelectrolyte (CPE) label having many redox-active sites and (ii) a readily prepared, thin polymeric monolayer-modified indium–tin oxide electrode. In addition, the long CPE label (∼18 nm for 10 kDa) can approach the electrode within the working distance after sandwich-type target-specific binding, and fast CPE-mediated oxidation of ammonia borane along the entire CPE backbone affords high signal amplification.</P><P><B>Graphic Abstract</B> <IMG SRC='http://pubs.acs.org/appl/literatum/publisher/achs/journals/content/jacsat/2018/jacsat.2018.140.issue-7/jacs.7b12382/production/images/medium/ja-2017-12382z_0006.gif'></P><P><A href='http://pubs.acs.org/doi/suppl/10.1021/ja7b12382'>ACS Electronic Supporting Info</A></P>