• Journal of Multimedia Information System
• /
• v.7 no.2
• /
• pp.141-146
• /
• 2020

Cine-Hangeul is a program that can predict the running time of a movie based on the screenplay before production. This paper seeks to verify the prediction reporting function of Cine-Hangeul, which is the standard Korean screenplay format. Moreover, this paper presents a method to increase the accuracy of the Cine-Hangeul reporting function. The objective of this paper is to offer a correction method based on scientific evidence because the current Cine-Hangeul reporting function has many errors. The verification process for five scenarios and movies confirmed that the default setting value of Cine- Hangeul's screening time prediction reporting was many errors. Cine-Hangeul analyzes the amount of textual information to predict the time of the scene and the time of the dialogue and helps predict the total time of the movie. Therefore, if a certain amount of text information is not available, the accuracy is unreliable. The current Cine-Hangeul prediction report confirms that the efficiency is high when the scenario volume is about 90 to 100 pages. As a result, prediction of screening time by Cine-Hangeul, a Korean scenario standard format program, confirmed the verification that it could secure the same level of reliability as the actual screening time by correcting the reporting settings. This verification also affirms that when applying about 50 percent of the basic set of screening time reporting, it is almost identical to the screening time.

• Nuclear Engineering and Technology
• /
• v.41 no.8
• /
• pp.1045-1052
• /
• 2009

A calculation model is developed to predict the transient free surface flow on the containment floor following a loss-of-coolant accident (LOCA) of pressurized water reactors (PWR) for the use of debris transport evaluation. The model solves the two-dimensional Shallow Water Equation (SWE) using a finite volume method (FVM) with unstructured triangular meshes. The numerical scheme is based on a fully explicit predictor-corrector method to achieve a fast-running capability and numerical accuracy. The Harten-Lax-van Leer (HLL) scheme is used to reserve a shock-capturing capability in determining the convective flux term at the cell interface where the dry-to-wet changing proceeds. An experiment simulating a sudden break of a water reservoir with L-shape open channel is calculated for validation of the present model. It is shown that the present model agrees well with the experiment data, thus it can be justified for the free surface flow with accuracy. From the calculation of flow field over the simplified containment floor of APR1400, the important phenomena of free surface flow including propagations and interactions of waves generated by local water level distribution and reflection with a solid wall are found and the transient flow rates entering the Holdup Volume Tank (HVT) are obtained within a practical computational resource.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2004.11a
• /
• pp.244-249
• /
• 2004

There have been many attempts to predict the secondary structure content of a protein from its primary sequence, which serves as the first step in a series of bioinformatics processes to gain knowledge of the structure and function of a protein. Most of them assumed that prediction relying on the information of the amino acid composition of a protein can be successful. Several approaches expanded the amount of information by including the pair amino acid composition of two adjacent residues. Recent methods achieved a remarkable improvement in prediction accuracy by using this expanded composition information. The overall average errors of two successful methods were 6.1% and 3.4%. This work was motivated by the observation that evolutionarily related proteins share the similar structure. After manipulating the values of the frequency matrix obtained by running PSI-BLAST, inputs of an artificial neural network were constructed by taking the ratio of the amino acid composition of the evolutionarily related proteins with a query protein to the background probability. Although we did not utilize the expanded composition information of amino acid pairs, we obtained the comparable accuracy, with the overall average error being 3.6%.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.29 no.10
• /
• pp.811-816
• /
• 2018

When applying the MLFMM algorithm to a large scattering problem, the accuracy of the calculation of the transfer function has a crucial effect on the final simulation results. The numerical accuracy for the double integral on the unit sphere is strongly dependent on the sampling number. With an increasing the sampling points, the overall required memory and running time of the MLFMM simulation also increases. Hence, an optimal over-sampling rate for the number of the sampling points is numerically obtained, which is verified for a real large scattering problem.

• 한국전산유체공학회:학술대회논문집
• /
• 2011.05a
• /
• pp.327-330
• /
• 2011

CFD simulation is widely used in various industries, universities and research centers. In Korea most of the researchers use foreign commercial S/W packages especially in industries. But commercial CFD packages have some problems as limit to source code and very high license foe. So from several years ago open source CFD code has been widely spread as an alternative. But in Korea there are a few users of open source code. Insufficiency of performance validation as for accuracy, robustness, convenience and parallel speed-up is important obstacles of open source code. So we tested some validation cases as to incompressible external aerodynamics and internal flaws and now are doing compressible flaws. As the first stage of compressible flow validation, we simulated Korea next generation high speed train(HEMU). It's running condition is 400km/hr and maximum Mach number reaches up to 0.4. With the high speed train we tested accuracy, robustness and parallel performance of open source CFD code OpenFOAM Because there isn't experimental data we compared results with widely used commercial code. When use $1^{st}$ order upwind scheme aerodynamic forces are very similar to commercial code. But using $2^{nd}$ order upwind scheme there was some discrepancy. The reason of the difference is not clear yet. Mesh manipulation, domain decomposition, post-processing and robustness are satisfactory. Paralle lperformance is similar to commercial code.

• Journal of Korea Multimedia Society
• /
• v.17 no.4
• /
• pp.506-513
• /
• 2014

In recent years, research on user's activity recognition using a smart phone has attracted a lot of attentions. A smart phone has various sensors, such as camera, GPS, accelerometer, audio, etc. In addition, smart phones are carried by many people throughout the day. Therefore, we can collect log data from smart phone sensors. The log data can be used to analyze user activities. This paper proposes an approach to inferring a user's physical activities based on the tri-axis accelerometer of smart phone. We propose recognition method for four activity which is physical activity; sitting, standing, walking, running. We have to convert accelerometer raw data so that we can extract features to categorize activities. This paper introduces a recognition method that is able to high detection accuracy for physical activity modes. Using the method, we developed an application system to recognize the user's physical activity mode in real-time. As a result, we obtained accuracy of over 80%.

• Smart Media Journal
• /
• v.9 no.2
• /
• pp.92-98
• /
• 2020

This paper proposes a vehicle license plate recognition system using light weight deep learning models without high-end server. The proposed license plate recognition system consists of 3 steps: [license plate detection]-[character area segmentation]-[character recognition]. SSD-Mobilenet was used for license plate detection, ResNet with localization was used for character area segmentation, ResNet was used for character recognition. Experiemnts using Samsung Galaxy S7 and LG Q9, accuracy showed 85.3% accuracy and around 1.1 second running time.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.2
• /
• pp.63-70
• /
• 2018

Safety management agent manages the risk behavior of the worker with the naked eye, but there is a real difficulty for one the agent to manage all the workers. In this paper, IoT device is attached to a harness safety belt that a worker wears to solve this problem, and behavior data is upload to the cloud in real time. We analyze the upload data through the deep learning and analyze the risk behavior of the worker. When the analysis result is judged to be dangerous behavior, we designed and implemented a system that informs the manager through monitoring application. In order to confirm that the risk behavior analysis through the deep learning is normally performed, the data values of 4 behaviors (walking, running, standing and sitting) were collected from IMU sensor for 60 minutes and learned through Tensorflow, Inception model. In order to verify the accuracy of the proposed system, we conducted inference experiments five times for each of the four behaviors, and confirmed the accuracy of the inference result to be 96.0%.

• Journal of Korea Multimedia Society
• /
• v.25 no.1
• /
• pp.39-51
• /
• 2022

Image-based gender classification and age estimation of human are classic problems in computer vision. Most of researches in this field focus just only one task of either gender classification or age estimation and most of the reported methods for each task focus on accuracy performance and are not computationally light. Thus, running both tasks together simultaneously on low cost mobile or embedded systems with limited cpu processing speed and memory capacity are practically prohibited. In this paper, we propose a novel light-weight gender classification and age estimation method based on ensemble multitasking deep learning with light-weight processing neural network architecture, which processes both gender classification and age estimation simultaneously and in real-time even for embedded systems. Through experiments over various well-known datasets, it is shown that the proposed method performs comparably to the state-of-the-art gender classification and/or age estimation methods with respect to accuracy and runs fast enough (average 14fps) on a Jestson Nano embedded board.

• Korean Journal of Community Nutrition
• /
• v.22 no.5
• /
• pp.413-425
• /
• 2017

Objectives: Accurate assessment of energy expenditure is important for estimation of energy requirements in athletic children. The objective of this study was to evaluate the accuracy of accelerometer for prediction of selected activities' energy expenditure and intensity in athletic elementary school children. Methods: The present study involved 31 soccer players (16 males and 15 females) from an elementary school (9-12 years). During the measurements, children performed eight selected activities while simultaneously wearing the accelerometer and carrying the portable indirect calorimeter. Five equations (Freedson/Trost, Treuth, Pate, Puyau, Mattocks) were assessed for the prediction of energy expenditure from accelerometer counts, while Evenson equation was added for prediction of activity intensity, making six equations in total. The accuracy of accelerometer for energy prediction was assessed by comparing measured and predicted values, using the paired t-test. The intensity classification accuracy was evaluated with kappa statistics and ROC-Curve. Results: For activities of lying down, television viewing and reading, Freedson/Trost, Treuth were accurate in predicting energy expenditure. Regarding Pate, it was accurate for vacuuming and slow treadmill walking energy prediction. Mattocks was accurate in treadmill running activities. Concerning activity intensity classification accuracy, Pate (kappa=0.72) had the best performance across the four intensities (sedentary, light, moderate, vigorous). In case of the sedentary activities, all equations had a good prediction accuracy, while with light activities and Vigorous activities, Pate had an excellent accuracy (ROC-AUC=0.91, 0.94). For Moderate activities, all equations showed a poor performance. Conclusions: In conclusion, none of the assessed equations was accurate in predicting energy expenditure across all assessed activities in athletic children. For activity intensity classification, Pate had the best prediction accuracy.