• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.107-119
• /
• 2023

The technology that recognizes a soldier's motion and movement status has recently attracted large attention as a combination of wearable technology and artificial intelligence, which is expected to upend the paradigm of troop management. The accuracy of state determination should be maintained at a high-end level to make sure of the expected vital functions both in a training situation; an evaluation and solution provision for each individual's motion, and in a combat situation; overall enhancement in managing troops. However, when input data is given as a timer series or sequence, existing feedforward networks would show overt limitations in maximizing classification performance. Since human behavior data (3-axis accelerations and 3-axis angular velocities) handled for military motion recognition requires the process of analyzing its time-dependent characteristics, this study proposes a high-performance data-driven classifier which utilizes the long-short term memory to identify the order dependence of acquired data, learning to classify eight representative military operations (Sitting, Standing, Walking, Running, Ascending, Descending, Low Crawl, and High Crawl). Since the accuracy is highly dependent on a network's learning conditions and variables, manual adjustment may neither be cost-effective nor guarantee optimal results during learning. Therefore, in this study, we optimized hyperparameters using Bayesian optimization for maximized generalization performance. As a result, the final architecture could reduce the error rate by 62.56% compared to the existing network with a similar number of learnable parameters, with the final accuracy of 98.39% for various military operations.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.8
• /
• pp.67-75
• /
• 2023

Stock price prediction is an important topic extensively discussed in the financial market, but it is considered a challenging subject due to numerous factors that can influence it. In this research, performance was compared and analyzed by applying time series prediction models (LSTM, GRU) and non-time series prediction models (RF, SVR, KNN, LGBM) that do not take into account the temporal dependence of data into stock price prediction. In addition, various data such as stock price data, technical indicators, financial statements indicators, buy sell indicators, short selling, and foreign indicators were combined to find optimal predictors and analyze major factors affecting stock price prediction by industry. Through the hyperparameter optimization process, the process of improving the prediction performance for each algorithm was also conducted to analyze the factors affecting the performance. As a result of feature selection and hyperparameter optimization, it was found that the forecast accuracy of the time series prediction algorithm GRU and LSTM+GRU was the highest.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.6
• /
• pp.1171-1180
• /
• 2019

Emotion is a psycho-physiological process that plays an important role in human interactions. Affective computing is centered on the development of human-aware artificial intelligence that can understand and regulate emotions. This field of study is also critical as mental diseases such as depression, autism, attention deficit hyperactivity disorder, and game addiction are associated with emotion. Despite the efforts in emotions recognition and emotion detection from nonstationary, detecting emotions from abnormal EEG signals requires sophisticated learning algorithms because they require a high level of abstraction. In this paper, we investigated LSTM hyperparameters for an optimal emotion EEG classification. Results of several experiments are hereby presented. From the results, optimal LSTM hyperparameter configuration was achieved.

• Journal of the Korean Society of Radiology
• /
• v.17 no.5
• /
• pp.709-715
• /
• 2023

In this study, the performance evaluation of image quality for noise reduction was implemented using the U-net deep learning architecture in computed tomography (CT) images. In order to generate input data, the Gaussian noise was applied to ground truth (GT) data, and datasets were consisted of 8:1:1 ratio of train, validation, and test sets among 1300 CT images. The Adagrad, Adam, and AdamW were used as optimizer function, and 10, 50 and 100 times for number of epochs were applied. In addition, learning rates of 0.01, 0.001, and 0.0001 were applied using the U-net deep learning model to compare the output image quality. To analyze the quantitative values, the peak signal to noise ratio (PSNR) and coefficient of variation (COV) were calculated. Based on the results, deep learning model was useful for noise reduction. We suggested that optimized hyper parameters for noise reduction in CT images were AdamW optimizer function, 100 times number of epochs and 0.0001 learning rates.

• Korean Journal of Remote Sensing
• /
• v.36 no.6_2
• /
• pp.1591-1604
• /
• 2020

This study is a study on classifying land cover by applying high-resolution satellite images to deep learning algorithms and verifying the performance of algorithms for each spatial object. For this, the Fully Convolutional Network-based algorithm was selected, and a dataset was constructed using Kompasat-3 satellite images, land cover maps, and forest maps. By applying the constructed data set to the algorithm, each optimal hyperparameter was calculated. Final classification was performed after hyperparameter optimization, and the overall accuracy of DeeplabV3+ was calculated the highest at 81.7%. However, when looking at the accuracy of each category, SegNet showed the best performance in roads and buildings, and U-Net showed the highest accuracy in hardwood trees and discussion items. In the case of Deeplab V3+, it performed better than the other two models in fields, facility cultivation, and grassland. Through the results, the limitations of applying one algorithm for land cover classification were confirmed, and if an appropriate algorithm for each spatial object is applied in the future, it is expected that high quality land cover classification results can be produced.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.307-307
• /
• 2022

최근 기후변화와 댐 상류 토지이용 변화 등과 같은 다양한 원인에 의해 댐 유입량의 변동성이 증가하면서 댐 관리 및 운영조작 의사 결정에 어려움이 발생하고 있다. 따라서 이러한 댐 유입량의 변동 특성을 반영하여 댐 유입량을 정확하고 효율적으로 예측할 수 있는 방안이 필요한 실정이다. 머신러닝 기술이 발전하면서 Auto-ML(Automated Machine Learning)이 다양한 분야에서 활용되고 있다. Auto-ML은 데이터 전처리, 최적 알고리즘 선택, 하이퍼파라미터 튜닝, 모델 학습 및 평가 등의 모든 과정을 자동화하는 기술이다. 그러나 아직까지 수문 분야에서 댐 유입량을 예측하기 위한 모델을 개발하는데 있어서 Auto-ML을 활용한 사례는 부족하고, 특히 댐 유입량의 예측 정확성을 확보하기 위해 High-inflow and low-inflow 의 변동 특성을 고려한 하이브리드 결합 방식을 통해 Auto-ML 기반 앙상블 모델을 개발하고 평가한 연구는 없다. 본 연구에서는 Auto-ML의 패키지 중 Auto-sklearn을 통해 홍수기, 비홍수기 유입량 변동 특성을 반영한 하이브리드 앙상블 댐 유입량 예측 모델을 개발하였다. 소양강댐을 대상으로 적용한 결과, 하이브리드 Auto-sklearn 앙상블 모델의 댐 유입량 예측 성능은 R² 0.868, RMSE 66.23 m³/s, MAE 16.45 m³/s로 단일 Auto-sklearn을 통해 구축 된 앙상블 모델보다 전반적으로 우수한 것으로 나타났다. 특히 FDC (Flow Duration Curve)의 저수기, 갈수기 구간에서 두 모델의 유입량 예측 경향은 큰 차이를 보였으며, 하이브리드 Auto-sklearn 모델의 예측 값이 관측 값과 더욱 유사한 것으로 나타났다. 이는 홍수기, 비홍수기 구간에 대한 앙상블 모델이 독립적으로 구축되는 과정에서 각 모델에 대한 하이퍼파라미터가 최적화되었기 때문이라 판단된다. 향후 본 연구의 방법론은 보다 정확한 댐 유입량 예측 자료를 생성하기 위한 방안 수립뿐만 아니라 다양한 분야의 불균형한 데이터셋을 이용한 앙상블 모델을 구축하는데도 유용하게 활용될 수 있을 것으로 사료된다.

• Korean Journal of Remote Sensing
• /
• v.34 no.5
• /
• pp.811-827
• /
• 2018

The purpose of this study is to compare machine learning algorithm and deep learning algorithm in crop classification using multi-temporal remote sensing data. For this, impacts of machine learning and deep learning algorithms on (a) hyper-parameter and (2) training sample size were compared and analyzed for Haenam-gun, Korea and Illinois State, USA. In the comparison experiment, support vector machine (SVM) was applied as machine learning algorithm and convolutional neural network (CNN) was applied as deep learning algorithm. In particular, 2D-CNN considering 2-dimensional spatial information and 3D-CNN with extended time dimension from 2D-CNN were applied as CNN. As a result of the experiment, it was found that the hyper-parameter values of CNN, considering various hyper-parameter, defined in the two study areas were similar compared with SVM. Based on this result, although it takes much time to optimize the model in CNN, it is considered that it is possible to apply transfer learning that can extend optimized CNN model to other regions. Then, in the experiment results with various training sample size, the impact of that on CNN was larger than SVM. In particular, this impact was exaggerated in Illinois State with heterogeneous spatial patterns. In addition, the lowest classification performance of 3D-CNN was presented in Illinois State, which is considered to be due to over-fitting as complexity of the model. That is, the classification performance was relatively degraded due to heterogeneous patterns and noise effect of input data, although the training accuracy of 3D-CNN model was high. This result simply that a proper classification algorithms should be selected considering spatial characteristics of study areas. Also, a large amount of training samples is necessary to guarantee higher classification performance in CNN, particularly in 3D-CNN.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.5
• /
• pp.331-337
• /
• 2023

In this paper, we propose a deep learning model to detect lumbar 3 (L3) CT images to determine the occurrence and degree of sarcopenia. In addition, we would like to propose an optimization technique that uses oversampling ratio and class weight as design parameters to address the problem of performance degradation due to data imbalance between L3 level and non-L3 level portions of CT data. In order to train and test the model, a total of 150 whole-body CT images of 104 prostate cancer patients and 46 bladder cancer patients who visited Gangneung Asan Medical Center were used. The deep learning model used ResNet50, and the design parameters of the optimization technique were selected as six types of model hyperparameters, data augmentation ratio, and class weight. It was confirmed that the proposed optimization-based L3 level extraction model reduced the median L3 error by about 1.0 slices compared to the control model (a model that optimized only 5 types of hyperparameters). Through the results of this study, accurate L3 slice detection was possible, and additionally, we were able to present the possibility of effectively solving the data imbalance problem through oversampling through data augmentation and class weight adjustment.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.07a
• /
• pp.53-56
• /
• 2022

본 논문에서는 심전도(ECG) 센서와 에지 컴퓨팅(Edge computing)을 활용하여 실시간 데이터와 Bayesian optimization을 통한 기계학습 알고리즘으로 재활 로봇에서 발목을 제어할 수 있는 Parameter(외골격 관련) 최적값을 출력한다. 심전도 센서 적용을 기반으로 하는 바이오 데이터 기술, 기계 학습(Bayesian optimization) 모델 접근 방식과 하드웨어 결합으로 재활 로봇 모터를 제어할 수 있는 Parameter 제공과 실시간 모터 제어 운영할 수 있도록 분석 플랫폼을 구축한다. 이 플랫폼을 이용해보다 효과적인 이동형 로봇설계 및 처리 방법을 연결할 수 있는 발판을 마련하였고, 로봇제어에 많이 사용하고 있는 매트랩 시뮬링크(Matlab simulink)를 연결할 수 있는 범용 통신 지원한다. 센서-전처리-인공지능 알고리즘-모터 제어 Parameter로 연계되는 데이터 가공과 처리 방법으로 최근 분석 기법을 적용하여 바이오 데이터 연구 활동과 이동형 재활 로봇 관련 데이터 분석 분야를 쉽게 접근할 수 있도록 한다.

• Journal of the Korean Society of Radiology
• /
• v.18 no.1
• /
• pp.21-26
• /
• 2024

The one of the famous deep learning models for object detection task is you only look once version 5 (YOLOv5) framework based on the one stage architecture. In addition, YOLOv5 model indicated high performance for accurate lesion detection using the bottleneck CSP layer and skip connection function. The purpose of this study was to evaluate the performance of YOLOv5 framework according to various hyperparameters in position emission tomogrpahy (PET) phantom images. The dataset was obtained from QIN PET segmentation challenge in 500 slices. We set the bounding box to generate ground truth dataset using labelImg software. The hyperparameters for network train were applied by changing optimization function (SDG, Adam, and AdamW), activation function (SiLU, LeakyRelu, Mish, and Hardwish), and YOLOv5 model size (nano, small, large, and xlarge). The intersection over union (IOU) method was used for performance evaluation. As a results, the condition of outstanding performance is to apply AdamW, Hardwish, and nano size for optimization function, activation function and model version, respectively. In conclusion, we confirmed the usefulness of YOLOv5 network for object detection performance in nuclear medicine images.