Daegu Gyeongbuk Institute of Science and Technology
DGIST Library Institutional RepositoryNot a member yet
12664 research outputs found
Sort by
Acidic CO2 Electroreduction for High CO2 Utilization: Catalysts, Electrodes, and Electrolyzers
No full textThe electrochemical carbon dioxide (CO2) reduction reaction (CO2RR) is considered a promising technology for converting atmospheric CO2 into value-added compounds by utilizing renewable energy. The CO2RR has developed in various ways over the past few decades, including product selectivity, current density, and catalytic stability. However, its commercialization is still unsuitable in terms of economic feasibility. One of the major challenges in its commercialization is the low single-pass conversion efficiency (SPCE) of CO2, which is primarily caused by the formation of carbonate (CO32−) in neutral and alkaline electrolytes. Notably, the majority of CO2RRs take place in such media, necessitating significant energy input for CO2 regeneration. Therefore, performing the CO2RR under conditions that minimize CO32− formation to suppress reactant and electrolyte ion loss is regarded an optimal strategy for practical applications. Here, we introduce the recent progress and perspectives in the electrochemical CO2RR in acidic electrolytes, which receives great attention because of the inhibition of CO32− formation. This includes the categories of nanoscale catalytic design, microscale microenvironmental effects, and bulk scale applications in electrolyzers for zero carbon loss reactions. Additionally, we offer insights into the issue of limited catalytic durability, a notable drawback under acidic conditions and propose guidelines for further development of the acidic CO2RR. © 2024 The Royal Society of Chemistry.FALSEsciescopu
Biomimetic Fe7S8/Carbon electrocatalyst from [FeFe]-Hydrogenase for improving pH-Universal electrocatalytic hydrogen production
No full textEfficient and cost-effective electrocatalysts that can operate across a wide range of pH conditions are essential for green hydrogen production. Inspired by biological systems, Fe7S8 nanoparticles incorporated on polydopamine matrix electrocatalyst were synthesized by co-precipitation and annealing process. The resulting Fe7S8/C electrocatalyst possesses a three-dimensional structure and exhibits enhanced electrocatalytic performance for hydrogen production across various pH conditions. Notably, the Fe7S8/C electrocatalyst demonstrates exceptional activity, achieving low overpotentials of 90.6, 45.9, and 107.4mV in acidic, neutral, and alkaline environments, respectively. Electrochemical impedance spectroscopy reveals that Fe7S8/C exhibits the lowest charge transfer resistance under neutral conditions, indicating an improved proton-coupled electron transfer process. Continuous-wave electron paramagnetic resonance results confirm a change in the valence state of Fe from 3+ to 1+ during the hydrogen evolution reaction (HER). These findings closely resemble the behavior of natural [FeFe]-hydrogenase, known for its superior hydrogen production in neutral conditions. The remarkable performance of our Fe7S8/C electrocatalyst opens up new possibilities for utilizing bioinspired materials as catalysts for the HER. © 2023 The Authors. Aggregate published by SCUT, AIEI, and John Wiley & Sons Australia, Ltd.TRU
Effects of Polybutylene Succinate Content on the Rheological Properties of Polylactic Acid/Polybutylene Succinate Blends and the Characteristics of Their Fibers
Get PDFPolylactic acid (PLA) and polybutylene succinate (PBS) are gaining prominence as environmentally friendly alternatives to petroleum-based polymers due to their inherent biodegradability. For their textile applications, this research is focused on exploring the effects of PBS content on the rheological properties of PLA/PBS blends and the characteristics of PLA/PBS blend fibers. PLA/PBS blends and fibers with varying PBS contents (0 to 10 wt.%) were prepared using melt-blending and spinning methods. Uniform morphologies of the PLA/PBS blends indicated that PBS was compatible with PLA, except at 10% PBS content, where phase separation occurred. The introduction of PBS reduced the complex viscosity of the blends, influencing fiber properties. Notably, PLA/PBS fibers with 7% PBS exhibited improved crystallinity, orientation factor, and elasticity (~16.58%), with a similar tensile strength to PLA fiber (~3.58 MPa). The results suggest that an optimal amount of PBS enhances alignment along the drawing direction and improves the molecular motion in PLA/PBS blend fiber. This study highlights the potential of strategically blending PBS to improve PLA fiber characteristics, promising advancement in textile applications. © 2024 by the authors. Licensee MDPI, Basel, Switzerland.FALS
An Electronic apparatus, Face Recognition system and Method for preventing spoofing thereof
No full text전자 장치 및 얼굴 인식 시스템, 그리고 이의 스푸핑 방지 방법이 제공된다. 본 전자 장치의 스푸핑 판단 방법은, 카메라로부터 획득된 이미지를 입력받는 단계, 입력된 이미지에 포함된 사용자의 얼굴 영역을 검출하는 단계, 검출된 사용자 얼굴 영역에 기초하여 사용자를 인식하는 단계, 이미지를 분석하여 이미지가 스푸핑 이미지(spoofing image)인지 여부를 판단하는 단계, 상기 사용자가 수행할 태스크를 요청하는 메시지를 제공하는 단계, 상기 사용자가 상기 태스크를 수행하는지 여부를 판단하여 스푸핑 여부를 판단하는 단계 및 상기 스푸핑 이미지인지 여부를 판단한 결과와 스푸핑 여부를 판단한 결과에 기초하여 사용자를 인증하는 단계를 포함한다
Improving 3D Human Modeling Accuracy Using 2D Image Segmentation
No full text본 연구에서는 2D 이미지 기반의 신체 실루엣을 활용하여 더욱 정확한 3D 신체 모델링을 위한 새로운 접근법을 제안한다. 기존 연구에서 다양한 헤어스타일로 인해 발생하는 머리 영역의 불확실성을 줄이고자 측면 이미지에서 머리 부분을 제거하고 정면 이미지는 그대로 보존하였다. 하지만, 이러한 접근법은 신체 측면에서 팔과 몸통이 겹쳐지는 문제를 일으켜, 정확한 신체 크기 측정에 한계를 나타낸다. 이에 따라 본 연구에서는 이러한 문제를 해결하기 위해 신체 각 부분에 대해 세분화된 세그멘테이션 레이블링을 도입하였고, 이를 통해 3D 모델의 정확성을 향상시키고자 하였다. 특히, 팔, 몸통, 다리의 각 부분을 독립적으로 구분하여 3D 복원 성능을 대폭 개선하는 방안을 제시하였다
Deep Learning-Based Hash Function Cryptanalysis
No full textThis paper analyzes the strength of Message Digest Algorithm (MD5) by performing deep learning-based Encryption Emulation (EE) and Plaintext Recovery (PR) attacks. We convert randomly generated S12-bit arrays, messages, into 128-bit arrays, digests, with MD5 in different numbers of steps. Furthermore, two different structures of deep learning models, fully-connected neural network and Bidirectional Long Short-Term Memory (BiLSTM), are used in attacks and trained to analyze MD5 automatically. As a result, the BiLSTM shows better prediction accuracy than the fully-connected neural network. Moreover, the PR attack is more challenging than the EE attack. © 2024 IEEE
Multi-Patching: Life-Log Classification with the Reconstructed Representation of Multivariate Time Series
No full textUnderstanding human beings requires analyzing human behavior with their life-log data. The life-log data is typically represented as multivariate time series data. This data is characterized by its extensive volume and the unforeseen emergence of missing values. These characteristics make the analysis challenging. This paper tackles these difficulties by generating reconstructed representations of the life-log data to utilize an LSTM autoencoder. Our method captures and compresses short-term patterns occurring within a single cycle. These reconstructed life-log data are fed into multivariate time series classification (MTSC) backbone models. This method not only improves performance but also efficiently manages memory usage. Experimental results showed that memory usage was reduced by 88.01 % while performance increased by 1.73 %. Achieving a weighted sum of F1-score of 5.91 on the test dataset confirmed the model's effectiveness. © 2024 IEEE
Analysis on Degree of Symmetry in Both Legs with Walking Motion using a Compact Radar Sensor
No full textFor this article, we analyze degree of symmetry in both legs with walking motion of an individual by using a compact continuous waveform Doppler radar. Generally, time-varying Doppler frequencies of both legs in a spectrogram includes two main parameters such as step length and radial upper leg length. Thus, the degree of symmetry in both legs can be presented as a function consisting of main parameters, so it can lead to efficient non-contact walking motion examination of patients undergoing rehabilitation training. In experiments, the degree of symmetry in both legs of an individual was efficiently analyzed from several walking motions. © 2024 IEEE
Landing-Type Aware Multi-Drone Route Generation for Last-Mile Delivery Service
No full textWe consider the problem of generating delivery routes for multiple drones in the last-mile delivery service. In particular, the landing type - how a parcel is to be dropped off from a drone - is explicitly modeled in terms of the landing area and the landing time, which was not considered in other drone delivery works. A Mixed Integer Linear Programming (MILP) problem is formulated to optimize the delivery route for each drone by minimizing the total delivery completion time. Our preliminary result shows that landing types affect the total delivery completion time significantly, even with a small number of drones. Therefore, it is necessary to explicitly consider the characteristics of landing types for more realistic delivery route generation of a large number of drones. © 2024 IEEE
Integrated Communication and Binary State Detection from Hoeffding's Perspective
No full textThis work considers a problem of integrated sensing and communication (ISAC) from an information-theoretic per-spective, in which the goal of sensing is to detect a binary state. We assume a broadcast channel that consists of a transmitter, a communication receiver, and a binary detector where the detector's channel is affected by an unknown binary state. Unlike most approaches where the detection error probability is considered, in our work, we analyze and disaggregate the error probability into more detailed components, false alarm and missed detection error exponents. We fully characterize the optimal three-way tradeoff between the coding rate for communication and the two possibly nonidentical error exponents for sensing in the asymptotic regime. The achievability and converse proofs rely on the analysis of the cumulant-generating function of the log-likelihood ratio. We also provide numerical evaluations of our results for binary and Gaussian channels. © 2024 International Federation for Information Processing - IFIP