• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.75-78
• /
• 2023

본 논문은 객체탐지 모델 중 주류를 이루고 있는 YOLO의 v1부터 v8까지의 특성을 비교 분석하여 각각의 버전에 최적화할 수 있는 모델에 대한 연구이다. 연구 결과 v1, v2는 정확성이 최우선인 모델에 적합하다. 반면, v3, v4는 속도가 우선인 모델에 적합하다. 또한 v5, v6는 정확도와 속도 사이의 균형이 필요한 모델에 적합하다는 결론을 얻었다. v7, v8은 메모리 및 컴퓨팅 성능에 제약이 있는 경우 주로 적용이 가능하며, 적은 연산과 저 메모리 사용으로 객체를 탐지하여 포즈추정이나 객체 추적 등을 적용할 모델에 적합하다는 결과를 확인하였다.

• Journal of the Korean Geotechnical Society
• /
• v.38 no.8
• /
• pp.53-66
• /
• 2022

The national ground survey database (GeoInfo) distributes numerous ground survey data nationwide. Many standard penetration test results exist in this database; however, the number of shear wave velocity (V_S) data is small. Hence, to use abundant standard penetration test-N values to predict V_S, this study proposed a new empirical N-V_S relationship model using GeoInfo data. The proposed N-V_S model is a single equation regardless of geological layer types; the layer type only specifies the upper limit of V_S. To validate the proposed model, residual analysis was performed using a test dataset that was not used for the model development. Therefore, this study's proposed model performed better than N-V_S models from previous studies. Since the N-V_S model in this study was developed using sufficient data from GeoInfo, we expect that it is the most applicable to GeoInfo dataset for V_S prediction.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.3
• /
• pp.98-110
• /
• 2018

In this paper, we propose both Dense grid model and Anchor model to improve the recognition rate of dynamic objects. Two experiments are conducted to study the performance of two proposed CNNs models (Dense grid model and Anchor model), which are to detect dynamic objects. In the first experiment, YOLO-v2 network is adjusted, and then fine-tuned on KITTI datasets. The Dense grid model and Anchor model are then compared with YOLO-v2. Regarding to the evaluation, the two models outperform YOLO-v2 from 6.26% to 10.99% on car detection at different difficulty levels. In the second experiment, this paper conducted further training of the models on a new dataset. The two models outperform YOLO-v2 up to 22.40% on car detection at different difficulty levels.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.47 no.9
• /
• pp.24-29
• /
• 2010

This paper presents the RF modeling for silicon nanowire MOSFET with 30 nm channel length and 5 nm channel radius. Equations for analytical parameter extraction are derived by analysis of Y-parameter. Accuracies of the new model and extracted parameters have been verified by 3-dimensional device simulation data up to 100 GHz. The model verifications are performed under conditions of saturation region ($V_{gs}$ = $_{ds}$ = 1 V) and linear region ($V_{gs}$ = 1 V, $V_{ds}$ = 0.5 V). The RMS modeling error of Y-parameters was calculated to be 1 %.

• Journal of Broadcast Engineering
• /
• v.28 no.3
• /
• pp.267-274
• /
• 2023

There is a problem that the size of the dataset is insufficient due to the limitation of the collection of the medical image public data, so there is a possibility that the existing studies are overfitted to the public dataset. In this paper, we compare the performance of eight (Unet, X-Net, HarDNet, SegNet, PSPNet, SwinUnet, 3D-ResU-Net, UNETR) medical image semantic segmentation models to revalidate the superiority of existing models. Anatomical Tracings of Lesions After Stroke (ATLAS) V1.2, a public dataset for stroke diagnosis, is used to compare the performance of the models and the performance of the models in ATLAS V2.0. Experimental results show that most models have similar performance in V1.2 and V2.0, but X-net and 3D-ResU-Net have higher performance in V1.2 datasets. These results can be interpreted that the models may be overfitted to V1.2.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.593-594
• /
• 2020

본 논문은 급격하게 발전하는 IT기술의 발전과 그로 인한 기업들의 경쟁 속에서 수 많은 프로젝트들이 생성되는 현대 사회에서 이런 프로젝트들이 경쟁에서 이길 수 있도록 프로젝트가 빠르고 완성도 있게 만들어질 필요를 느끼고 기존의 프로젝트 개발 방법론과 테스트 설계 방법론 중 V모델을 연구해서 프로젝트 관리 시스템에 등록된 데이터를 기반으로 테스트가 자동 설계될 수 있게 함으로써 프로젝트 전체 기간을 줄이고 프로젝트 비용을 줄일 수 있는 방법을 제안한다. 이 논문에서 제안하는 모델을 확장시켜서 프로젝트 테스트 관리에 불필요한 설계 시간과 테스트 관리에 수월할 것이라 기대한다.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.40 no.2
• /
• pp.75-83
• /
• 2016

Since a turbomachine has complex flow characteristics, which are caused by adverse pressure gradient and high speed motion, an elaborate turbulence model is needed to accurately predict the flow. Some turbulence models such as an algebraic or a two-equation eddy viscosity model have been used for in-house RANS-code, but it is difficult to obtain good result for several complex flows. In this study, Durbin's V2-F turbulence model, which has been known for better prediction for severe flow separation, is applied to T-Flow. It was validated for simple cases such as channel and compressor cascade, and its applicability to turbomachinery was shown by analyzing internal flow of a single rotor. As a result, the V2-F turbulence model shows better blade surface pressure distribution than the one-and-two equation turbulence model.

• Journal of the Korea Society for Simulation
• /
• v.24 no.3
• /
• pp.55-60
• /
• 2015

This paper presents an SBA (simulation-based acquisition) effectiveness analysis environment using the RUP (Relational Unified Process) model. The RUP model has 4 phases which cover the whole development steps such as requirement analysis, design, development and test. By applying the RUP model, SW development can be represented with the iterations of developments for each phase. Such a characteristics of the model would make the model suitable for defense acquisition. In this paper, we show the relation between the RUP model and V process model, which is the foundation for defense acquisition. In order to show how the model could be applied to SBA effectiveness analysis, graphical user interfaces for the analysis are presented at the end of the paper.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2017.10a
• /
• pp.454-456
• /
• 2017

Current and capacitance-voltage characteristics of tunnel field effect transistor (TFET) with various quantum models were investigated. Density gradient, Bohm quantum potential (BQP), and Vandort quantum correction are used with calibrating against Schrodinger-Poisson model. Drive-currents in all models. are decreased. When only BQP is used, SS and $V_{onset}$ are fixed but drive-current is decreased 3 times more than those of no quantum model. And When BQP with Vandort and density gradient are used, SS increased more than 40 mV./dec and $V_{onset}$ shifted as 0.07 eV.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.26 no.4
• /
• pp.419-428
• /
• 2020

The flow past a wall mounted square cylinder, a typical and basic shape of building, bridge or offshore structure, was simulated using URANS computation through adoption of three turbulence models, namely, the k-ε model, k-ω model, and the v²-f model. It is well known that this flow is naturally unstable due to the Karman vortex shedding and exhibits a complex flow structure in the wake region. The mean flow field including velocity profiles and the dominant frequency of flow oscillation that was from the simulations discussed earlier were compared with the experimental data observed by Wang et al. (2004; 2006). Based on these comparisons it was found that the v²-f model is most accurate for the URANS simulation; moreover, the k-ω model is also acceptable. However, the k-ε model was found to be unsuitable in this case. Therefore, v²-f model is proved to be an excellent choice for the analysis of flow with massive separation. Therefore, it is expected to be used in future by studies aiming to control the flow separation.