• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.3
• /
• pp.134-139
• /
• 2023

Rehabilitators perform outpatient treatment and daily rehabilitation exercises to recover physical function with the aim of quickly returning to society after surgical treatment. Unlike performing exercises in a hospital with the help of a professional therapist, there are many difficulties in performing rehabilitation exercises by the patient on a daily basis. In this paper, we propose a CNN-LSTM-based upper limb rehabilitation real-time monitoring system so that patients can perform rehabilitation efficiently and with correct posture on a daily basis. The proposed system measures biological signals through shoulder-mounted hardware equipped with EMG and IMU, performs preprocessing and normalization for learning, and uses them as a learning dataset. The implemented model consists of three polling layers of three synthetic stacks for feature detection and two LSTM layers for classification, and we were able to confirm a learning result of 97.44% on the validation data. After that, we conducted a comparative evaluation with the Teachable machine, and as a result of the comparative evaluation, we confirmed that the model was implemented at 93.6% and the Teachable machine at 94.4%, and both models showed similar classification performance.

• Smart Media Journal
• /
• v.12 no.11
• /
• pp.48-56
• /
• 2023

The technology of Three-dimensional human posture estimation is used in sports, motion recognition, and special effects of video media. Among various methods for this, multi-view 3D human pose estimation is essential for precise estimation even in complex real-world environments. But Existing models for multi-view 3D human posture estimation have the disadvantage of high order of time complexity as they use 3D feature maps. This paper proposes a method to extend an existing monocular viewpoint multi-frame model based on Transformer with lower time complexity to 3D human posture estimation for multi-viewpoints. To expand to multi-viewpoints our proposed method first generates an 8-dimensional joint coordinate that connects 2-dimensional joint coordinates for 17 joints at 4-vieiwpoints acquired using the 2-dimensional human posture detector, CPN(Cascaded Pyramid Network). This paper then converts them into 17×32 data with patch embedding, and enters the data into a transformer model, finally. Consequently, the MLP(Multi-Layer Perceptron) block that outputs the 3D-human posture simultaneously updates the 3D human posture estimation for 4-viewpoints at every iteration. Compared to Zheng[5]'s method the number of model parameters of the proposed method was 48.9%, MPJPE(Mean Per Joint Position Error) was reduced by 20.6 mm (43.8%) and the average learning time per epoch was more than 20 times faster.

• Nuclear Engineering and Technology
• /
• v.55 no.12
• /
• pp.4357-4366
• /
• 2023

MARS-KS, a domestic regulatory confirmatory code of Republic of Korea, had been developed by integrating RELAP5/MOD2 and COBRA-TF. The integration of COBRA-TF allowed to extend the capability of MARS-KS, limited to one-dimensional analysis, to multi-dimensional analysis. The use of COBRA-TF was mainly focused on subchannel analyses for simulating multi-dimensional behavior within the reactor core. However, this feature has been remained as a legacy without ongoing maintenance. Meanwhile, MARS-KS also includes its own multidimensional component, namely MULTID, which is also feasible to simulate three-dimensional convection and diffusion. The MULTID is capable of modeling the turbulent diffusion using simple mixing length model. The implementation of the turbulent mixing is of importance for analyzing the reactor core where a disturbing cross-sectional structure of rod bundle makes the flow perturbation and corresponding mixing stronger. In addition, the presence of this turbulent behavior allows the secondary transports with net mass exchange between subchannels. However, a series of assessments performed in previous studies revealed that the turbulence model of the MULTID could not simulate the aforementioned effective mixing occurred in the subchannel-scale problems. This is obvious consequence since the physical models of the MULTID neglect the effect of mass transport and thereby, it cannot model the void drift effect and resulting phasic distribution within a bundle. Thus, in this study, the turbulence mixing model of the MULTID has been improved by means of the inter-channel mixing model, widely utilized in subchannel analysis, in order to extend the application of the MULTID to small-scale problems. A series of assessments has been performed against rod bundle experiments, namely GE 3X3 and PSBT, to evaluate the performance of the introduced mixing model. The assessment results revealed that the application of the inter-channel mixing model allowed to enhance the prediction of the MULTID in subchannel scale problems. In addition, it was indicated that the code could not predict appropriate phasic distribution in the rod bundle without the model. Considering that the proper prediction of the phasic distribution is important when considering pin-based and/or assembly-based expressions of the reactor core, the results of this study clearly indicate that the inter-channel mixing model is required for analyzing the rod bundle, appropriately.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.4
• /
• pp.178-185
• /
• 2023

Rapid urbanization and advancements in technology have led to a surge in the number of automobiles, resulting in frequent traffic accidents, and consequently, an increase in human casualties and economic losses. Therefore, there is a need for technology that can predict the risk of traffic accidents to prevent them and minimize the damage caused by them. Traffic accidents occur due to various factors including traffic congestion, the traffic environment, and road conditions. These factors give traffic accidents spatiotemporal characteristics. This paper analyzes traffic accident data to understand the main characteristics of traffic accidents and reconstructs the data in a time series format. Additionally, an LSTM-MLP based model that excellently captures spatiotemporal characteristics was developed and utilized for traffic accident prediction. Experiments have proven that the proposed model is more rational and accurate in predicting the risk of traffic accidents compared to existing models. The traffic accident risk prediction model suggested in this paper can be applied to systems capable of real-time monitoring of road conditions and environments, such as navigation systems. It is expected to enhance the safety of road users and minimize the social costs associated with traffic accidents.

• Journal of Intelligence and Information Systems
• /
• v.17 no.4
• /
• pp.241-254
• /
• 2011

Financial time-series forecasting is one of the most important issues because it is essential for the risk management of financial institutions. Therefore, researchers have tried to forecast financial time-series using various data mining techniques such as regression, artificial neural networks, decision trees, k-nearest neighbor etc. Recently, support vector machines (SVMs) are popularly applied to this research area because they have advantages that they don't require huge training data and have low possibility of overfitting. However, a user must determine several design factors by heuristics in order to use SVM. For example, the selection of appropriate kernel function and its parameters and proper feature subset selection are major design factors of SVM. Other than these factors, the proper selection of instance subset may also improve the forecasting performance of SVM by eliminating irrelevant and distorting training instances. Nonetheless, there have been few studies that have applied instance selection to SVM, especially in the domain of stock market prediction. Instance selection tries to choose proper instance subsets from original training data. It may be considered as a method of knowledge refinement and it maintains the instance-base. This study proposes the novel instance selection algorithm for SVMs. The proposed technique in this study uses genetic algorithm (GA) to optimize instance selection process with parameter optimization simultaneously. We call the model as ISVM (SVM with Instance selection) in this study. Experiments on stock market data are implemented using ISVM. In this study, the GA searches for optimal or near-optimal values of kernel parameters and relevant instances for SVMs. This study needs two sets of parameters in chromosomes in GA setting : The codes for kernel parameters and for instance selection. For the controlling parameters of the GA search, the population size is set at 50 organisms and the value of the crossover rate is set at 0.7 while the mutation rate is 0.1. As the stopping condition, 50 generations are permitted. The application data used in this study consists of technical indicators and the direction of change in the daily Korea stock price index (KOSPI). The total number of samples is 2218 trading days. We separate the whole data into three subsets as training, test, hold-out data set. The number of data in each subset is 1056, 581, 581 respectively. This study compares ISVM to several comparative models including logistic regression (logit), backpropagation neural networks (ANN), nearest neighbor (1-NN), conventional SVM (SVM) and SVM with the optimized parameters (PSVM). In especial, PSVM uses optimized kernel parameters by the genetic algorithm. The experimental results show that ISVM outperforms 1-NN by 15.32%, ANN by 6.89%, Logit and SVM by 5.34%, and PSVM by 4.82% for the holdout data. For ISVM, only 556 data from 1056 original training data are used to produce the result. In addition, the two-sample test for proportions is used to examine whether ISVM significantly outperforms other comparative models. The results indicate that ISVM outperforms ANN and 1-NN at the 1% statistical significance level. In addition, ISVM performs better than Logit, SVM and PSVM at the 5% statistical significance level.

• Journal of Korean Society of Transportation
• /
• v.22 no.3 s.74
• /
• pp.145-157
• /
• 2004

The cycle length design model of the Korean traffic responsive signal control systems is devised to vary a cycle length as a response to changes in traffic demand in real time by utilizing parameters specified by a system operator and such field information as degrees of saturation of through phases. Since no explicit guideline is provided to a system operator, the system tends to include ambiguity in terms of the system optimization. In addition, the cycle lengths produced by the existing model have yet been verified if they are comparable to the ones minimizing delay. This paper presents the studies conducted (1) to find shortcomings embedded in the existing model by comparing the cycle lengths produced by the model against the ones minimizing delay and (2) to propose a new direction to design a cycle length minimizing delay and excluding such operator oriented parameters. It was found from the study that the cycle lengths from the existing model fail to minimize delay and promote intersection operational conditions to be unsatisfied when traffic volume is low, due to the feature of the changed target operational volume-to-capacity ratio embedded in the model. The 64 different neural network based cycle length design models were developed based on simulation data surrogating field data. The CORSIM optimal cycle lengths minimizing delay were found through the COST software developed for the study. COST searches for the CORSIM optimal cycle length minimizing delay with a heuristic searching method, a hybrid genetic algorithm. Among 64 models, the best one producing cycle lengths close enough to the optimal was selected through statistical tests. It was found from the verification test that the best model designs a cycle length as similar pattern to the ones minimizing delay. The cycle lengths from the proposed model are comparable to the ones from TRANSYT-7F.

• Economic and Environmental Geology
• /
• v.15 no.3
• /
• pp.123-154
• /
• 1982

Petrochemical, K-Ar dating, Sand Rb/Sr isotopes, metallogenic zoning, paleomagnetic and geotectonic studies of the Gyongsang basin were carried out to examine applicability of plate tectonics to the post-late Cretaceous igneous activity and metallogeny in the southeastern part of Korean Peninsula. The results obtained are as follows: 1. Bulgugsa granitic rocks range from granite to adamellite, whose Q-Ab-Or triangular diagram indicates that the depth and pressure at which the magma consolidated increase from coast to inland varying from 6 km, 0.5-3.3 kb in the coastal area to 17 km, 0.5-10 kb in the inland area. 2. The volcanic rocks in Gyongsang basin range from andesitic to basaltic rocks, and the basaltic rocks are generally tholeiitic in the coastal area and alkali basalt in the inland area. 3. The volcanic rocks of the area have the initial ratio of Sr^{87}/Sr^{86} varying from 0.706 to 0.707 which suggests a continental origin; the ratio of Rb/Sr changing from 0.079-0.157 in the coastal area to 0.021-0.034 in the inland area suggests that the volcanism is getting younger toward coastal side, which may indicate a retreat in stage of differentiation if they were derived from a same magma. The K_2O/SiO_2 (60%) increases from about 1.0 in the coastal area to about 3.0 in the inland area, which may suggest an increase indepth of the Benioff zone, if existed, toward inland side. 4. The K-Ar ages of volcanic rocks were measured to be 79.4 m.y. near Daegu, and 61.7 m.y. near Busan indicating a southeastward decrease in age. The ages of plutonic rocks also decrease toward the same direction with 73 m.y. near Daegu, and 58 m.y. near Busan, so that the volcanism predated the plutonism by 6 m.y. in the continental interior and 4 m.y. along the coast. Such igneous activities provide a positive evidence for an applicability of plate tectonics to this area. 5. Sulfur isotope analyses of sulfide minerals from 8 mines revealed that these deposits were genetically connected with the spacially associated ingeous rocks showing relatively narrow range of ${\delta}^{34}S$ values (-0.9‰ to +7.5‰ except for +13.3 from Mulgum Mine). A sequence of metallogenic zones from the coast to the inland is delineated to be in the order of Fe-Cu zone, Cu-Pb-Zn zone, and W-Mo zone. A few porphyry type copper deposits are found in the Fe-Cu zone. These two facts enable the sequence to be comparable with that of Andean type in South America. 6. The VGP's of Cretaceous and post Cretaceous rocks from Korea are located near the ones($71^{\circ}N$, $180^{\circ}E$ and $90^{\circ}N$, $110^{\circ}E$) obtained from continents of northern hemisphere. This suggests that the Korean peninsula has been stable tectonically since Cretaceous, belonging to the Eurasian continent. 7. Different polar wandering path between Korean peninsula and Japanese islands delineates that there has been some relative movement between them. 8. The variational feature of declination of NRM toward northwestern inland side from southeastern extremity of Korean peninsula suggests that the age of rocks becomes older toward inland side. 9. The geological structure(mainly faults) and trends of lineaments interpreted from the Landsat imagery reveal that NNE-, NWW- and NEE-trends are predominant in the decreasing order of intensity. 10. The NNE-trending structures were originated by tensional and/or compressional forces, the directions of which were parallel and perpendicular respectively to the subduction boundary of the Kula plate during about 90 m.y. B.P. The NWW-trending structures were originated as shear fractures by the same compressional forces. The NEE-trending structures are considered to be priginated as tension fractures parallel to the subduction boundary of the Kula plate during about 70 m.y. B.P. when Japanese islands had drifted toward southeast leaving the Sea of Japan behind. It was clearly demonstrated by many authors that the drifting of Japanese islands was accompanied with a rotational movement of a clock-wise direction, so that it is inferred that subduction boundary had changed from NNE- to NEE-direction. A number of facts and features mentioned above provide a suite of positive evidences enabling application of plate tectonics to the late Cretaceous-early Tertiary igneous activity and metallogeny in the area. Synthesizing these facts, an arc-trench system of continental margin-type is adopted by reconstructing paleogeographic models for the evolution of Korean peninsula and Japan islands. The models involve an extention mechanism behind the are(proto-Japan), by which proto-Japan as of northeastern continuation of Gyongsang zone has been drifted rotationally toward southeast. The zone of igneous activity has also been migrated from the inland in late-Cretaceous to the peninsula margin and southwestern Japan in Tertiary.

• Journal of The Korean Association For Science Education
• /
• v.21 no.2
• /
• pp.357-384
• /
• 2001

In this paper, we described practical teaching-learning plans based on three different theoretical approaches to Integrated Science Education (ISE): a knowledge centered ISE, a social problem centered ISE, and an individual interest centered ISE. We believe that science teachers can understand integrated science education through this paper and they are able to apply simultaneously our integrated science teaching materials to their real instruction in classroom. For this we developed integrated science teaching-learning plans for the topic of energy which has a integrated feature strongly among integrated science subject contents. These modules were based upon the teaching strategies of 'Energy' following each integrated directions organized in the previous paper (Three Strategies for Integrated Science Teaching of "Energy" Applying Knowledge, Social Problem, and Individual Interest Centered Approaches) and we applied instruction models fitting each features of integrated directions to the teaching strategies of 'Energy'. There is a concrete describing on the above three integrated science teaching-learning plans as follows. 1. For the knowledge centered integration, we selected the topic, 'Journey of Energy' and we tried to integrate the knowledge of physics, chemistry, biology, and earth science applying the instruction model of 'Free Discovery Learning' which is emphasized on concepts and inquiry. 2. For the social problem centered integration, we selected the topic, 'Future of Energy' to resolve the science-related social problems and we applied the instruction model of 'Project Learning' which is emphasized on learner's cognitive process to the topic. 3. For the individual interest centered integration, we selected the topic, 'Transformation of Energy' for the integration of science and individual interest and we applied the instruction model of 'Project Learning' centering learner's interest and concern. Based upon the above direction, we developed the integrated science teaching-learning plans as following steps. First, we organized 'Integrated Teaching-Learning Contents' according to the topics. Second, based upon the above organization, we designed 'Instructional procedures' to integrate within the topics. Third, in accordance with the above 'Instructional Procedures', we created 'Instructional Coaching Plan' that can be applied in the practical world of real classrooms. These plans can be used as models for the further development of integrated science instruction for teacher preparation, textbook development, and classroom learning.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.26 no.2
• /
• pp.69-84
• /
• 1984

This study was carried out to investigate the time distribution of single storms and to establish the model of storm patterns in korea. Rainfall recording charts collected from 42 metheorological stations covering the Korean peninsula were analyzed. A single storm was defined as a rain period seperated from preceding and succeeding rainfall by 6 hours and more. Among the defined single storms, 1199 storms exceeding total rainfall of 80 mm were qualified for the study. Storm patterns were cklassified by quartile classification method and the relationship between cummulative percent of rainfalls and cummulative storm time was established for each quartile storm group. Time distribution models for each stations were prepared through the various analytical and inferential procedures. Obtained results are summarized as follows: 1. The percentile frequency of quartile storms for the first to the fourth quartile were 22.0%, 26.5%, 28.9% and 22.6%, respectively. The large variation of percentile frequency was show between the same quartile storms. The advanced type storm pattern was predominant in the west coastal type storm patterns predominantly when compared to the single storms with small total rainfalls. 3. The single storms with long storm durations tended to show delayed type storm patterns predominantly when compared to the single storms with short storm durations. 4. The percentile time distribution of quartile storms for 42 rin gaging stations was estimated. Large variations were observed between the percentiles of time distributions of different stations. 5. No significant differences were generally found between the time distribution of rainfalls with greater total rainfall and with less total rainfall. This fact suggests that the size of the total rainfall of single storms was not the main factor affecting the time distribution of heavy storms. 6. Also, no significant difference were found between the time distribution of rainfalls with long duration and with short duration. The fact indicates that the storm duration was no the main factor affecting the time distribution of heavy storms. 7. In Korea, among all single storms, 39.0% show 80 to 100mm of total rainfall which stands for the mode of the frequency distribution of total rainfalls. The median value of rainfalls for all single storms from the 42 stations was 108mm. The shape of the frequency distribution of total rainfalls showed right skewed features. No significant differences were shown in the shape of distribution histograms for total rainfall of quartile storms. The mode of rainfalls for the advanced type quartile storms was 80~100mm and their frequencies were 39~43% for respective quartiles. For the delayed type quartile storms, the mode was 80~100mm and their frequencies were 36!38%. 8. In Korea, 29% of all single storms show 720 to 1080 minutes of storm durations which was the highest frequency in the frequency distribution of storm durations. The median of the storm duration for all single storms form 42 stations was 1026 minutes. The shape of the frequency distribution was right skewed feature. For the advanced type storms, the higher frequency of occurrence was shown by the single storms with short durations, whereas for the delayed type quartile storms, the higher frequency was shown gy the long duration single storms. 9. The total rainfall of single storms was positively correlated to storm durations in all the stations throughout the nation. This fact was also true for most of the quartile storms. 10. The third order polynomial regression models were established for estimating the time distribution of quartile storms at different stations. The model test by relative error method resulted good agreements between estimated and observed values with the relative error of less than 0.10 in average.

• Journal of Intelligence and Information Systems
• /
• v.24 no.3
• /
• pp.1-19
• /
• 2018

Large amount of data is now available for research and business sectors to extract knowledge from it. This data can be in the form of unstructured data such as audio, text, and image data and can be analyzed by deep learning methodology. Deep learning is now widely used for various estimation, classification, and prediction problems. Especially, fashion business adopts deep learning techniques for apparel recognition, apparel search and retrieval engine, and automatic product recommendation. The core model of these applications is the image classification using Convolutional Neural Networks (CNN). CNN is made up of neurons which learn parameters such as weights while inputs come through and reach outputs. CNN has layer structure which is best suited for image classification as it is comprised of convolutional layer for generating feature maps, pooling layer for reducing the dimensionality of feature maps, and fully-connected layer for classifying the extracted features. However, most of the classification models have been trained using online product image, which is taken under controlled situation such as apparel image itself or professional model wearing apparel. This image may not be an effective way to train the classification model considering the situation when one might want to classify street fashion image or walking image, which is taken in uncontrolled situation and involves people's movement and unexpected pose. Therefore, we propose to train the model with runway apparel image dataset which captures mobility. This will allow the classification model to be trained with far more variable data and enhance the adaptation with diverse query image. To achieve both convergence and generalization of the model, we apply Transfer Learning on our training network. As Transfer Learning in CNN is composed of pre-training and fine-tuning stages, we divide the training step into two. First, we pre-train our architecture with large-scale dataset, ImageNet dataset, which consists of 1.2 million images with 1000 categories including animals, plants, activities, materials, instrumentations, scenes, and foods. We use GoogLeNet for our main architecture as it has achieved great accuracy with efficiency in ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Second, we fine-tune the network with our own runway image dataset. For the runway image dataset, we could not find any previously and publicly made dataset, so we collect the dataset from Google Image Search attaining 2426 images of 32 major fashion brands including Anna Molinari, Balenciaga, Balmain, Brioni, Burberry, Celine, Chanel, Chloe, Christian Dior, Cividini, Dolce and Gabbana, Emilio Pucci, Ermenegildo, Fendi, Giuliana Teso, Gucci, Issey Miyake, Kenzo, Leonard, Louis Vuitton, Marc Jacobs, Marni, Max Mara, Missoni, Moschino, Ralph Lauren, Roberto Cavalli, Sonia Rykiel, Stella McCartney, Valentino, Versace, and Yve Saint Laurent. We perform 10-folded experiments to consider the random generation of training data, and our proposed model has achieved accuracy of 67.2% on final test. Our research suggests several advantages over previous related studies as to our best knowledge, there haven't been any previous studies which trained the network for apparel image classification based on runway image dataset. We suggest the idea of training model with image capturing all the possible postures, which is denoted as mobility, by using our own runway apparel image dataset. Moreover, by applying Transfer Learning and using checkpoint and parameters provided by Tensorflow Slim, we could save time spent on training the classification model as taking 6 minutes per experiment to train the classifier. This model can be used in many business applications where the query image can be runway image, product image, or street fashion image. To be specific, runway query image can be used for mobile application service during fashion week to facilitate brand search, street style query image can be classified during fashion editorial task to classify and label the brand or style, and website query image can be processed by e-commerce multi-complex service providing item information or recommending similar item.