• Korean Journal of Remote Sensing
• /
• v.39 no.5_4
• /
• pp.1111-1123
• /
• 2023

In recent times, semantic image segmentation methods using deep learning models have been widely used for monitoring changes in surface attributes using remote sensing imagery. To enhance the performance of various UNet-based deep learning models, including the prominent UNet model, it is imperative to have a sufficiently large training dataset. However, enlarging the training dataset not only escalates the hardware requirements for processing but also significantly increases the time required for training. To address these issues, transfer learning is used as an effective approach, enabling performance improvement of models even in the absence of massive training datasets. In this paper we present three transfer learning models, UNet-ResNet50, UNet-VGG19, and CBAM-DRUNet-VGG19, which are combined with the representative pretrained models of VGG19 model and ResNet50 model. We applied these models to building extraction tasks and analyzed the accuracy improvements resulting from the application of transfer learning. Considering the substantial impact of learning rate on the performance of deep learning models, we also analyzed performance variations of each model based on different learning rate settings. We employed three datasets, namely Kompsat-3A dataset, WHU dataset, and INRIA dataset for evaluating the performance of building extraction results. The average accuracy improvements for the three dataset types, in comparison to the UNet model, were 5.1% for the UNet-ResNet50 model, while both UNet-VGG19 and CBAM-DRUNet-VGG19 models achieved a 7.2% improvement.

• The Journal of Korean Physical Therapy
• /
• v.27 no.2
• /
• pp.112-117
• /
• 2015

Purpose: The purpose of this study was to examine the effects of a task-specific obstacle crossing rehabilitation program on functional gait ability in patients with cerebellar ataxia. Overall, we sought to provide ataxia-specific locomotor rehabilitation guidelines for use in clinical practice based on quantitative evidence using relevant analysis of gait kinematics including valid clinical tests. Methods: Patients with cerebellar disease (n=13) participated in obstacle crossing training focusing on maintenance of dynamic balance and posture, stable transferring of body weight, and production of coordinated limb movements for 8 weeks, 2 times per week, 90 minutes per session. Throughout the training of body weight transfer, the instructions emphasized conscious perception and control of the center of body stability, trunk and limb alignment, and stepping kinematics during the practice of each walking phase. Results: According to the results, compared with pre-training data, foot clearance, pre-&post-obstacle distance, delay time, and total obstacle crossing time were increased after intervention. In addition, body COM measures indicated that body sway and movement variability, therefore posture stability during obstacle crossing, showed improvement after training. Based on these results, body sway was reduced and stepping pattern became more consistent during obstacle crossing gait after participation in patients with cerebellar ataxia. Conclusion: Findings of this study suggest that task-relevant obstacle crossing training may have a beneficial effect on recovery of functional gait ability in patients with cerebellar disease.

• Nuclear Engineering and Technology
• /
• v.55 no.11
• /
• pp.4102-4111
• /
• 2023

In order to better perform thermal hydraulic calculation and analysis of supercritical water reactor, based on the experimental data of supercritical water, the model training and predictive analysis of the heat transfer coefficient of supercritical water were carried out by using the support vector machine (SVM) algorithm. The changes in the prediction accuracy of the supercritical water heat transfer coefficient are analyzed by the changes of the regularization penalty parameter C, the slack variable epsilon and the Gaussian kernel function parameter gamma. The predicted value of the SVM model obtained after parameter optimization and the actual experimental test data are analyzed for data verification. The research results show that: the normalization of the data has a great influence on the prediction results. The slack variable has a relatively small influence on the accuracy change range of the predicted heat transfer coefficient. The change of gamma has the greatest impact on the accuracy of the heat transfer coefficient. Compared with the calculation results of traditional empirical formula methods, the trained algorithm model using SVM has smaller average error and standard deviations. Using the SVM trained algorithm model, the heat transfer coefficient of supercritical water can be effectively predicted and analyzed.

• Journal of the Semiconductor & Display Technology
• /
• v.20 no.1
• /
• pp.64-69
• /
• 2021

In this paper, a method for improving the defect classification performance in low contrast, ununiformity and featureless steel plate surfaces has been studied based on deep convolution neural network and transfer-learning neural network. The steel plate surface images have low contrast, ununiformity, and featureless, so that the contrast between defect and defect-free regions are not discriminated. These characteristics make it difficult to extract the feature of the surface defect image. A classifier based on a deep convolution neural network is constructed to extract features automatically for effective classification of images with these characteristics. As results of the experiment, AlexNet-based transfer-learning classifier showed excellent classification performance of 99.43% with less than 160 seconds of training time. The proposed classification system showed excellent classification performance for low contrast, ununiformity, and featureless surface images.

• The KSFM Journal of Fluid Machinery
• /
• v.10 no.3 s.42
• /
• pp.39-46
• /
• 2007

This study presents a numerical procedure to optimize the shape of staggered dimple surface to enhance turbulent heat transfer in a rectangular channel. The RBNN method is used as an optimization technique with Reynolds-averaged Navier-Stokes analysis of fluid flow and heat transfer with shear stress transport (SST) turbulence model. The dimple depth-to-dimple print diameter (d/D), channel height-to-dimple print diameter ratio (H/D), and dimple print diameter-to-pitch ratio (D/S) are chosen as design variables. The objective function is defined as a linear combination of heat transfer related term and friction loss related term with a weighting factor. Latin Hypercube Sampling (LHS) is used to determine the training points as a mean of the design of experiment. The optimum shape shows remarkable performance in comparison with a reference shape.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.2
• /
• pp.420-437
• /
• 2024

There are some problems in network traffic classification (NTC), such as complicated statistical features and insufficient training samples, which may cause poor classification effect. A NTC architecture based on one-dimensional Convolutional Neural Network (CNN) and transfer learning is proposed to tackle these problems and improve the fine-grained classification performance. The key points of the proposed architecture include: (1) Model classification--by extracting normalized rate feature set from original data, plus existing statistical features to optimize the CNN NTC model. (2) To apply transfer learning in the classification to improve NTC performance. We collect two typical network flows data from Youku and YouTube, and verify the proposed method through extensive experiments. The results show that compared with existing methods, our method could improve the classification accuracy by around 3-5%for Youku, and by about 7 to 27% for YouTube.

• Journal of Computing Science and Engineering
• /
• v.5 no.3
• /
• pp.257-268
• /
• 2011

Machine learning and data mining have found many applications in biological domains, where we look to build predictive models based on labeled training data. However, in practice, high quality labeled data is scarce, and to label new data incurs high costs. Transfer and multitask learning offer an attractive alternative, by allowing useful knowledge to be extracted and transferred from data in auxiliary domains helps counter the lack of data problem in the target domain. In this article, we survey recent advances in transfer and multitask learning for bioinformatics applications. In particular, we survey several key bioinformatics application areas, including sequence classification, gene expression data analysis, biological network reconstruction and biomedical applications.

• The Journal of Korean Physical Therapy
• /
• v.25 no.5
• /
• pp.273-281
• /
• 2013

Purpose: The purpose of this study is to determine the effect of motor imagery training on residual upper extremity strength and activities of daily living of chronic cervical spinal cord injury patients. Methods: Twelve ASIA A B patients, who had more than a 12-month duration of illness and C5 or 6 motor nerve injury level, were randomly divided into experimental group (n=6) and control group (n=6). Patients in the experimental group performed motor imagery training for five minutes prior to general muscle strengthening training, while those in the control group performed general muscle strengthening training only. The training was performed five times per week, 30 minutes per day, for a period of four weeks. General muscle strengthening training consisted of a progressive resistive exercise for residual upper extremity. Motor imagery training consisted of imagining this task performance. Before and after the training, EMG activity using BTS Pocket Electromyography and Spinal Cord Independent Measure III(SCIM III) were compared and analyzed. Results: The residual upper extremity muscle strengths showed improvement in both groups after training. Comparison of muscle strength improvement between the two groups showed a statistically significant improvement in the experimental group compared to the control group (p<0.05). SCIM III measurements showed significant improvement in the scores for Self-care and Transfer items in the experimental group. Conclusion: Motor imagery training was more effective than general muscle strengthening training in improving the residual upper extremity muscle strength and activities of daily living of patients with chronic cervical spinal cord injury.

• 한국전산유체공학회:학술대회논문집
• /
• 2008.03a
• /
• pp.229-232
• /
• 2008

This study presents a numerical procedure to optimize the shape of stepped circular pin-fins to enhance turbulent heat transfer. The KRG method is used as an optimization technique with Reynolds-averaged Navier-Stokes analysis of fluid flow and heat transfer with shear stress transport turbulent model. The objective function is defined as a linear combination of heat transfer and friction loss related terms with a weighting factor. Ten training points are obtained by Latin Hypercube Sampling for two design variables. Optimum shape has been successfully obtained with the increased objective function.

• 한국전산유체공학회:학술대회논문집
• /
• 2008.10a
• /
• pp.229-232
• /
• 2008

This study presents a numerical procedure to optimize the shape of stepped circular pin-fins to enhance turbulent heat transfer. The KRG method is used as an optimization technique with Reynolds-averaged Navier-Stokes analysis of fluid flow and heat transfer with shear stress transport turbulent model. The objective function is defined as a linear combination of heat transfer and friction loss related terms with a weighting factor. Ten training points are obtained by Latin Hypercube Sampling for two design variables. Optimum shape has been successfully obtained with the increased objective function.