• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2004.10a
• /
• pp.411-418
• /
• 2004

Using a level set method we develop a shape optimization method applied to energy flow problems in steady state. The boundaries are implicitly represented by the level set function obtainable from the 'Hamilton-Jacobi type' equation with the 'Up-wind scheme.' The developed method defines a Lagrangian function for the constrained optimization. It minimizes a generalized compliance, satisfying the constraint of allowable volume through the variations of implicit boundary. During the optimization, the boundary velocity to integrate the Hamilton-Jacobi equation is obtained from the optimality condition for the Lagrangian function. Compared with the established topology optimization method, the developed one has no numerical instability such as checkerboard problems and easy representation of topological shape variations.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2014.04a
• /
• pp.741-744
• /
• 2014

디비피디아 온톨로지는 위키피디아에서 구조화된 데이터를 추출한 지식 베이스이다. 이러한 지식 베이스의 자동 증강은 웹을 구조화하는 속도를 증가시키는데 큰기여를 할 수 있다. 본 연구에서는 한국어 디비피디아를 기반으로 새로운 트리플을 입력받아 기존의 지식 베이스를 자동 증강시키는 시스템을 소개한다. 스키마를 자동 증강하는 두 가지 알고리즘은 최하위 레벨인 인스턴스가 지닌 프로퍼티, 즉 rdf-triple 단위에서 진행되었다. 알고리즘을 사용한 결과 첫째, 확률적 격상 방법을 통해 단계별로 입력받는 인스턴스와 하위 클래스의 프로퍼티를 이용하여 상위 클래스의 스키마가 정교해졌다. 둘째, 이를 바탕으로 타입 분류가 되어 있지 않았던 인스턴스들이 가장 가까운 타입에 자동 분류되었다. 지식 베이스가 정교해지면서 재분류된 인스턴스와 새로운 트리플셋을 바탕으로 두 가지 알고리즘은 반복적으로 작동하며, 한국어 디비피디아 지식 베이스의 자동 증강을 이루었다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.5
• /
• pp.7-14
• /
• 2021

Tomato crops are highly affected by tomato diseases, and if not prevented, a disease can cause severe losses for the agricultural economy. Therefore, there is a need for a system that quickly and accurately diagnoses various tomato diseases. In this paper, we propose a system that classifies nine diseases as well as healthy tomato plants by applying various pretrained deep learning-based CNN models trained on an ImageNet dataset. The tomato leaf image dataset obtained from PlantVillage is provided as input to ResNet, Xception, and DenseNet, which have deep learning-based CNN architectures. The proposed models were constructed by adding a top-level classifier to the basic CNN model, and they were trained by applying a 5-fold cross-validation strategy. All three of the proposed models were trained in two stages: transfer learning (which freezes the layers of the basic CNN model and then trains only the top-level classifiers), and fine-tuned learning (which sets the learning rate to a very small number and trains after unfreezing basic CNN layers). SGD, RMSprop, and Adam were applied as optimization algorithms. The experimental results show that the DenseNet CNN model to which the RMSprop algorithm was applied output the best results, with 98.63% accuracy.

• Journal of Korea Multimedia Society
• /
• v.15 no.9
• /
• pp.1075-1085
• /
• 2012

In clinical research using medical images, the image segmentation is one of the most important processes. Especially, the hippocampal atrophy is helpful for the clinical Alzheimer diagnosis as a specific marker of the progress of Alzheimer. In order to measure hippocampus volume exactly, segmentation of the hippocampus is essential. However, the hippocampus has some features like relatively low contrast, low signal-to-noise ratio, discreted boundary in MRI images, and these features make it difficult to segment hippocampus. To solve this problem, firstly, We selected region of interest from an experiment image, subtracted a original image from the negative image of the original image, enhanced contrast, and applied anisotropic diffusion filtering and gaussian filtering as preprocessing. Finally, We performed an image segmentation using two level set methods. Through a variety of approaches for the validation of proposed hippocampus segmentation method, We confirmed that our proposed method improved the rate and accuracy of the segmentation. Consequently, the proposed method is suitable for segmentation of the area which has similar features with the hippocampus. We believe that our method has great potential if successfully combined with other research findings.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.6
• /
• pp.19-28
• /
• 2021

Nowadays, interest in health care is increasing due to Coronavirus (COVID-19), and a lot of people are doing home training as there are more difficulties in using fitness centers and public facilities that are used together. In this paper, we propose a personalized exercise recommendation algorithm using personalized propensity information to provide more accurate and meaningful exercise recommendation to home training users. Thus, we classify the data according to the criteria for obesity with a k-nearest neighbor algorithm using personal information that can represent individuals, such as eating habits information and physical conditions. Furthermore, we differentiate the exercise dataset by the level of exercise activities. Based on the neighborhood information of each dataset, we provide personalized exercise recommendations to users through a dimensionality reduction algorithm (SVD) among model-based collaborative filtering methods. Therefore, we can solve the problem of data sparsity and scalability of memory-based collaborative filtering recommendation techniques and we verify the accuracy and performance of the proposed algorithms.

• Journal of the Korea Computer Graphics Society
• /
• v.29 no.4
• /
• pp.1-6
• /
• 2023

To enhance immersive experiences for metaverse environements, background music is often used. However, the background music is mostly pre-matched and repeated which might occur a distractive experience to users as it does not align well with rapidly changing user-interactive contents. Thus, we implemented a system to provide a more immersive metaverse conversation experience by 1) developing a regression neural network that extracts emotions from an utterance using KEMDy20, the Korean multimodal emotion dataset 2) selecting music corresponding to the extracted emotions from an utterance by the DEAM dataset where music is tagged with arousal-valence levels 3) combining it with a virtual space where users can have a real-time conversation with avatars.

• The KIPS Transactions:PartA
• /
• v.12A no.5 s.95
• /
• pp.451-460
• /
• 2005

We present a methodology to design energy-efficient data paths using FPGAs. Our methodology integrates domain specific modeling, coarse-grained performance evaluation, design space exploration, and low-level simulation to understand the tradeoffs between energy, latency, and area. The domain specific modeling technique defines a high-level model by identifying various components and parameters specific to a domain that affect the system-wide energy dissipation. A domain is a family of architectures and corresponding algorithms for a given application kernel. The high-level model also consists of functions for estimating energy, latency, and area that facilitate tradeoff analysis. Design space exploration(DSE) analyzes the design space defined by the domain and selects a set of designs. Low-level simulations are used for accurate performance estimation for the designs selected by the DSE and also for final design selection We illustrate our methodology using a family of architectures and algorithms for matrix multiplication. The designs identified by our methodology demonstrate tradeoffs among energy, latency, and area. We compare our designs with a vendor specified matrix multiplication kernel to demonstrate the effectiveness of our methodology. To illustrate the effectiveness of our methodology, we used average power density(E/AT), energy/(area x latency), as themetric for comparison. For various problem sizes, designs obtained using our methodology are on average $25\%$ superior with respect to the E/AT performance metric, compared with the state-of-the-art designs by Xilinx. We also discuss the implementation of our methodology using the MILAN framework.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.4
• /
• pp.11-19
• /
• 2021

An abnormal object refers to a person, an object, or a mechanical device that performs abnormal and unusual behavior and needs observation or supervision. In order to detect this through artificial intelligence algorithm without continuous human intervention, a method of observing the specificity of temporal features using optical flow technique is widely used. In this study, an abnormal situation is identified by learning an algorithm that translates an input image frame to an optical flow image using a Generative Adversarial Network (GAN). In particular, we propose a technique that improves the pre-processing process to exclude unnecessary outliers and the post-processing process to increase the accuracy of identification in the test dataset after learning to improve the performance of the model's abnormal behavior identification. UCSD Pedestrian and UMN Unusual Crowd Activity were used as training datasets to detect abnormal behavior. For the proposed method, the frame-level AUC 0.9450 and EER 0.1317 were shown in the UCSD Ped2 dataset, which shows performance improvement compared to the models in the previous studies.

• Journal of Korea Multimedia Society
• /
• v.20 no.9
• /
• pp.1491-1501
• /
• 2017

In this paper, we propose a method to measure respiration rate by dividing the respiration related region in depth image using level set method. In the conventional method, the respiration related region was separated using the pre-defined region designated by the user. We separate the respiration related region using level set method combining shape prior knowledge. Median filter and clipping are performed as a preprocessing method for noise reduction in the depth image. As a feasibility test, respiration activity was recorded using depth camera in various environments with arm movements or body movements during breathing. Respiration activity was also measured simultaneously using a chest belt to verify the accuracy of calculated respiration rate. Experimental results show that our proposed method shows good performance for respiration rate estimation in various situation compared with the conventional method.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.01a
• /
• pp.217-218
• /
• 2021

본 논문에서는 다층 퍼셉트론을 활용하여 League of Legends 게임의 승패를 예측하는 Deep Neural Network 프로그램을 설계하는 방법을 제안한다. 연구 방법으로 한국 서버의 챌린저 리그에서 행해진 약 26000 경기 데이터 셋을 분석하여, 경기 도중 15분 데이터 중 드래곤 처치 수, 챔피언 레벨, 정령, 타워 처치 수가 게임 결과에 유의미한 영향을 끼치는 것을 확인하였다. 모델 설계는 softmax 함수보다 sigmoid 함수를 사용했을 때 더 높은 정확도를 얻을 수 있었다. 실제 LOL의 프로 게임 16경기를 예측한 결과 93.75%의 정확도를 도출했다. 게임 평균시간이 34분인 것을 고려하였을 때, 게임 중반 정도에 게임의 승패를 예측할 수 있음이 증명되었다. 본 논문에서 설계한 이 프로그램은 전 세계 E-sports 프로리그의 승패예측과 프로팀의 유용한 훈련지표로 활용 가능하다고 사료된다.