• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.986-989
• /
• 2019

인공지능에 대한 연구가 계속되고 있지만 음악, 미술, 문학 등 창의력과 예술성을 요구하는 부문에서는 인공지능을 적용하기 어렵다. 하지만 그중 음악 부문에 인공지능을 접목시켜 기존에 없던 곡을 작곡해보고자 한다. LSTM 기반의 모델을 ABC Notation 악보 데이터를 활용하여 학습시켜 사용자들이 음악적 지식 없이도 새로운 음악을 작곡할 수 있도록 딥러닝을 기반으로 한 BGM 음원 작곡 서비스를 제안한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.512-513
• /
• 2023

CNN 기반 Deep Learning 분야에서 객체 탐지 정확도를 높이기 위해 모델의 많은 Parameter 가 사용된다. 많은 Parameter 를 사용하게 되면 최소 하드웨어 성능 요구치가 상승하고 처리속도도 감소한다는 문제가 있어, 최소한의 정확도 하락으로 Parameter 를 줄이기 위한 여러 Pruning 기법이 사용된다. 본 연구에서는 Neural Architecture Search(NAS) 기반 Channel Pruning 인 Artificial Bee Colony(ABC) 알고리즘을 사용하였고, 기존 NAS 기반 Channel Pruning 논문들이 Classification Task 에서만 실험한 것과 달리 Object Detection Task 에서도 NAS 기반 Channel Pruning 을 적용하여 기존 Uniform Pruning 과 비교할 때 파라미터 수 대비 정확도가 개선됨을 확인하였다.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.3
• /
• pp.199-216
• /
• 2024

This research examines the factors influencing consumer purchase intentions in e-commerce live streaming, set against the backdrop of the internet celebrity economy. The investigation serves as a pivotal inquiry into the dynamics of this economy, striving to uncover the extent of internet celebrities' influence, particularly in terms of their economic impact. Employing the Emotional Behavioral Cognitive (ABC) attitude theory and the Stimulus Organism Response (S-O-R) theory as foundational frameworks, this study scrutinizes internet celebrity live streaming sales. It incorporates direct observations and leverages existing scholarly work to devise a tailored measurement scale and questionnaire. From this, a research model and hypotheses are developed, leading to the establishment of an empirical model. This empirical model is instrumental in statistically analyzing how e-commerce live streaming, within the internet celebrity economy context, shapes consumer purchase intentions. By integrating theoretical insights and empirical findings, the research elucidates the strategic dimensions and consumer behavior aspects in digital commerce. It enhances understanding of how internet celebrity influence intersects with consumer purchasing processes. Overall, this study contributes to the academic discourse on digital marketing and consumer behavior, providing a nuanced perspective on the mechanisms through which internet celebrities affect e-commerce. It offers valuable implications for marketers, strategists, and policymakers aiming to navigate the complex landscape of the internet celebrity economy.

• Information and Communications Magazine
• /
• v.19 no.11
• /
• pp.85-108
• /
• 2002

향후 다양한 무선 통신 규격들의 통합 수용을 위한 SDR (Software Defined Radio) 기술이IMT-2000 이후의 4세대 이동 통신 시스템을 위한 핵심 기술로써 심각하게 고려되고 있다. 이에 부응하여 SDR기반의 멀티모드용 통신 시스템을 구성하기 위한 주요 기술로서 디지털 IF 기술에 대한 필요성이 급속도로 고조되고 있는 상황이다. 최근 ABC/DAC 및 범용 디지털 신호처리 소자들의 고속화 및 고성능화로 인해 If (Intermediate Frequency) 대역과 기저대역 신호들 간의 직접 디지털 변환의 구현이 현실화되고 있다. 사용자의 관점에서 국지적으로 상용화되고 있거나 장래에 출현할 다양한 이동 통신 시스템 규격들 및 이에 대해 사업자들에게 할당되는 주파수 대역들이 서로 다른 점을 고려할 때, 이종 시스템 혹은 사업자들에게 할당된 주파수 대역에 구애받지 않고 언제 어디서나 자유롭게 무선 채널을 엑세스하고 또한 특정 채널을 임의로 선택하기 위한 디지털 If기술의 실현이 필수적이다 이러한 SDR기반 디지털 If 기술은 소프트웨어적으로 재구성 가능한 하드웨어 구조를 요구하며, 특정 이동 통신 규격의 물리 계층만을 지원하는 무선 인터페이스가 아닌 다중이동 통신 모드를 지원할 수 있는 유연성이 가미된 채널화 알고리즘이 필요하게 된다. 따라서 디지털 If기술은 무선 인터페이스 처리 부분, 즉 주파수 상 하향 변환 및 채별 선택 조합을 용도에 맞게 단일의 하드웨어 플렛폼 상에서 고속 디지털 신호처리 알고리즘을 기반으로 동작하기 위한 기능을 필연적으로 요구한다. 본 논문에서는 향후 SDR 기반의 기지국 및 단말기 운영 및 구생 모델을 제시하며, 디지털 If에 대한 필요성 및 동작 원리, 그리고 요소 기능들에 대한 구체적인 동작 원리 및 디지털 If와 더불어 활용 가능한 기술에 대하여 논의한다.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.64 no.6
• /
• pp.900-911
• /
• 2015

In this paper, 3D face recognition model is designed by using Polynomial based RBFNN(Radial Basis Function Neural Network) and PNN(Polynomial Neural Network). Also recognition rate is performed by this model. In existing 2D face recognition model, the degradation of recognition rate may occur in external environments such as face features using a brightness of the video. So 3D face recognition is performed by using 3D scanner for improving disadvantage of 2D face recognition. In the preprocessing part, obtained 3D face images for the variation of each pose are changed as front image by using pose compensation. The depth data of face image shape is extracted by using Multiple point signature. And whole area of face depth information is obtained by using the tip of a nose as a reference point. Parameter optimization is carried out with the aid of both ABC(Artificial Bee Colony) and PSO(Particle Swarm Optimization) for effective training and recognition. Experimental data for face recognition is built up by the face images of students and researchers in IC&CI Lab of Suwon University. By using the images of 3D face extracted in IC&CI Lab. the performance of 3D face recognition is evaluated and compared according to two types of models as well as point signature method based on two kinds of depth data information.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.43 no.7
• /
• pp.601-608
• /
• 2015

The static aeroelastic analysis and optimization of flexible wings are conducted for steady state conditions while both aerodynamic and structural parameters can be used as optimization variables. The system of multidisciplinary design optimization as a robust methodology to couple commercial codes for a static aeroelastic optimization purpose to yield a convenient adaptation to engineering applications is developed. Aspect ratio, taper ratio, sweepback angle are chosen as optimization variables and the skin thickness of the wing. The real-coded adaptive range multi-objective genetic algorithm code, which represents the global multi-objective optimization algorithm, was used to control the optimization process. The support vector regression(SVR) is applied for optimization, in order to reduce the time of computation. For this multi-objective design optimization problem, numerical results show that several useful Pareto optimal designs exist for the flexible wing.

• The Korean Journal of Malacology
• /
• v.23 no.1
• /
• pp.31-38
• /
• 2007

Population ecological parameters and stock biomass of the purplish washington clam, Saxidomus purpuratus (Sowerby), in the southern coastal areas of Korea were determined based on the fishery data from the Divers Fisheries Cooperative and other available biological data. Instantaneous coefficient of total mortality (Z) of purplish Washington clam was estimated to be 0.7479/year. The estimated instantaneous coefficient of natural mortality (M) was 0.2012/year. From the values of Z and M, the instantaneous coefficient of fishing mortality coefficient (F) for the recent years was calculated to be 0.4578/year. The age of purplish Washington clam at its first capture ($t_c$) was 2.7506 year. Yield-per-recruit and spawning biomass-per-recruit were estimated under harvest strategies that bases on $F_{max}$, $F_{0.1}$, $F_{35%}$ and $F_{40%}$ were shown as 81.60 g, 61.68 g, 115.07 g and 131.51 g, respectively. The acceptable biological catch (ABC) was estimated to be about 1,404 metric tons.

• Journal of the korean academy of Pediatric Dentistry
• /
• v.26 no.2
• /
• pp.218-231
• /
• 1999

Pyramidal cells in the hippocampal CA area were recorded from and filled with neurobiotin in anesthetized rats. The extent of their dendrites and the electropharmacological properties of membrane as well as the effect before and after neurobiotin injection were examined. Pyramidal cells had a high resting membrane potential, a low input resistance, and a large amplitude action potential. A afterhyperpolarization was followed a single action potential. Most pyramidal cells did not display a spontaneous firing. Pyramidal cell displayed weak inward rectification and anodal break excitation in response to negative current injection into the cell. Membrane properties of recorded neurons before and after neurobiotin injection with consecutive current injection were compared. Some properties were significantly increased after labelling(P>0.05); the duration and amplitude of sustained AHP, input resistance, and the number of action potentials for simultaneous intra- and extracellular stimulations. Neurobiotin-filled neurons showed pyramidal morphology. Cells were generally bipolar dendrite processes ramifying in stratum lacunosum-moleculare, radiatum, and oriens.