• Structural Monitoring and Maintenance
• /
• v.7 no.4
• /
• pp.283-294
• /
• 2020

In this paper, an adaptive MEMD based modal identification technique for linear time-invariant systems is proposed employing multiple vibration measurements. Traditional empirical mode decomposition (EMD) suffers from mode-mixing during sifting operations to identify intrinsic mode functions (IMF). MEMD performs better in this context as it considers multi-channel data and projects them into a n-dimensional hypercube to evaluate the IMFs. Using this technique, modal parameters of the structural system are identified. It is observed that MEMD has superior performance compared to its traditional counterpart. However, it still suffers from mild mode-mixing in higher modes where the energy contents are low. To avoid this problem, an adaptive filtering scheme is proposed to decompose the interfering modes. The Proposed modified scheme is then applied to vibrations of a reinforced concrete road bridge. Results presented in this study show that the proposed MEMD based approach coupled with the filtering technique can effectively identify the parameters of the dominant modes present in the structural response with a significant level of accuracy.

• Journal of Construction Engineering and Project Management
• /
• v.5 no.4
• /
• pp.16-22
• /
• 2015

This paper presents an automated computer vision-based system to update BIM data by leveraging multi-modal visual data collected from existing buildings under inspection. Currently, visual inspections are conducted for building envelopes or mechanical systems, and auditors analyze energy-related contextual information to examine if their performance is maintained as expected by the design. By translating 3D surface thermal profiles into energy performance metrics such as actual R-values at point-level and by mapping such properties to the associated BIM elements using XML Document Object Model (DOM), the proposed method shortens the energy performance modeling gap between the architectural information in the as-designed BIM and the as-is building condition, which improve the reliability of building energy analysis. Several case studies were conducted to experimentally evaluate their impact on BIM-based energy analysis to calculate energy load. The experimental results on existing buildings show that (1) the point-level thermography-based thermal resistance measurement can be automatically matched with the associated BIM elements; and (2) their corresponding thermal properties are automatically updated in gbXML schema. This paper provides practitioners with insight to uncover the fundamentals of how multi-modal visual data can be used to improve the accuracy of building energy modeling for retrofit analysis. Open research challenges and lessons learned from real-world case studies are discussed in detail.

• International conference on construction engineering and project management
• /
• 2015.10a
• /
• pp.532-536
• /
• 2015

This paper presents an automated computer vision-based system to update BIM data by leveraging multi-modal visual data collected from existing buildings under inspection. Currently, visual inspections are conducted for building envelopes or mechanical systems, and auditors analyze energy-related contextual information to examine if their performance is maintained as expected by the design. By translating 3D surface thermal profiles into energy performance metrics such as actual R-values at point-level and by mapping such properties to the associated BIM elements using XML Document Object Model (DOM), the proposed method shortens the energy performance modeling gap between the architectural information in the as-designed BIM and the as-is building condition, which improve the reliability of building energy analysis. The experimental results on existing buildings show that (1) the point-level thermography-based thermal resistance measurement can be automatically matched with the associated BIM elements; and (2) their corresponding thermal properties are automatically updated in gbXML schema. This paper provides practitioners with insight to uncover the fundamentals of how multi-modal visual data can be used to improve the accuracy of building energy modeling for retrofit analysis. Open research challenges and lessons learned from real-world case studies are discussed in detail.

• Land and Housing Review
• /
• v.4 no.1
• /
• pp.89-98
• /
• 2013

In recent years, it was recognized that aurban policy paradigm of sustainable urban growth management and transportation policies have been shifted : from the era of automobile-oriented policy to the era of public transport policy. Against this backdrop, the introduction of the multi-modal transfer center is very consequential. Therefore, it is necessary for the introduction of wide-area and local-centric transfer facilities, as well as the center of the country-led national backbone transfer center. This study was applicable at the multi-modal transfer center plans to introduce guidelines to provide transit facilities, considering the regional and functional characteristics to classify the types of multi-modal transfer center. The final types of multi-modal transfer center were classified into six, and by considering the combination of the criteria to be classified. The multi-modal transfer center type classification based on the case analysis of the types of facilities at domestic and abroad. If the data of multi-modal transfer are accumulated continuously, can expect a more reliable type classification.

• Journal of KSNVE
• /
• v.8 no.1
• /
• pp.139-145
• /
• 1998

A great deal of effort has been invested in upgrading the performance and the efficiency of dynamic analysis of mechanical structures. Using experimental modal analysis(EMA) or finite element analysis(FEA) data of mechanical structures, the performance and efficiency can be effectively evaluated. In order to analyze complex structures such as automobiles and aircrafts, for the sake of computing efficiency, the dynamic substructuring techniques that allow to predict the dynamic behavior of a structure are widely used. Through linking a modal model obtained from EMA and an analytical model obtained from FEA, the best conditioned strucutres can be proposed. In this study, a new algorithm of substructre synthesis method, Multi-FRF synthesis method, is proposed to analyze a structure composed of many substructures.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.526-527
• /
• 2023

Speech emotion recognition(SER) is one of the interesting topics in the machine learning field. By developing multi-modal speech emotion recognition system, we can get numerous benefits. This paper explain about fusing BERT as the text recognizer and CNN as the speech recognizer to built a multi-modal SER system.

• Journal of the Society of Naval Architects of Korea
• /
• v.50 no.2
• /
• pp.69-78
• /
• 2013

This study is constructed to investigate the sloshing effect on the motions of a two-dimensional rectangular cylinder experimentally and numerically. The modes of motion under consideration are sway and roll, and also experimental cases are divided by two categories; 1-DoF roll motion and 2-DoF motion (Coupling sway and roll). It is found that the sway response is considerably affected by the motion of the fluid, particularly near the sloshing natural frequency, while the roll response changes comparatively small. The dominant mode of motion is analyzed for 2-DoF experiments as well. The measured data for 1-DoF motions is compared with numerical results obtained by the Multi-modal approach. The numerical schemes vary in detail with the number of dominant sloshing modes; i.e. there is a single dominant mode for the Single-dominant method, while the Model 2 method assumes that the first two modes are superior. For the roll motion, numerical results obtained by the two different methods are relatively in good agreement with the experiments, and these two results are similar in most wave frequency range. However, the discrepancies are apparent where the fluid motion is not governed by a single mode. But both of numerical methods over-predict the motion at the vicinity of the sloshing natural frequency. In order to correct the discrepancy, the modal damping needs to be investigated more precisely. Furthermore, another multi-modal approach, such as the Boussinesq-type method, seems to be required in the region of the intermediate liquid.

• The Journal of Society for e-Business Studies
• /
• v.21 no.1
• /
• pp.147-163
• /
• 2016

Gender information of a smart device user is essential to provide personalized services, and multi-modal data obtained from the device is useful for predicting the gender of the user. However, the method for utilizing each of the multi-modal data for gender prediction differs according to the characteristics of the data. Therefore, in this study, an ensemble method for predicting the gender of a smart device user by using three classifiers that have text, application, and acceleration data as inputs, respectively, is proposed. To alleviate privacy issues that occur when text data generated in a smart device are sent outside, a classification method which scans smart device text data only on the device and classifies the gender of the user by matching text data with predefined sets of word. An application based classifier assigns gender labels to executed applications and predicts gender of the user by comparing the label ratio. Acceleration data is used with Support Vector Machine to classify user gender. The proposed method was evaluated by using the actual smart device log data collected from an Android application. The experimental results showed that the proposed method outperformed the compared methods.

• Journal of the Korean Society of Laryngology, Phoniatrics and Logopedics
• /
• v.13 no.2
• /
• pp.132-138
• /
• 2002

Background and Objectives : The complex physiologic structure of the larynx can vibrate in three or more different ways that yield acuostically and perceptually distinct vocal quality. The purpose of this study is to examine the normal range of voice parameters in Multi-Channel Phonatory Function Analyzer and investigate the difference of voice parameters according to the phonatory patterns. Materials and Methods : Forty normal adult speakers (20 men and 20 women) with age ranging from third to forth decades pronounce low, comfortable, and high tone /a/ ; comfortable tone /${\ae}$/, /i/, /o/, and /u/ : fry, falsetto. Voice was analyzed by Newly developed multi-channel phonatory function analyzer. Results : The normal range of voice parameters in this system was similar to the existing data. Fry shows high jitter and falsetto low SQ. Fry and falsetto show low OQ in men but no difference in women. Jitter, OQ and SQ were different between men and women in modal register, whereas there was no gender difference in fry and falsetto. In frequency magnitude spectrum and EGG, modal register, fry and falsetto have distinguishing pattern. Conclusions : Modal register, fry and falsetto are distinguishable in voice parameters and show different vibratory patterns.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.3
• /
• pp.670-684
• /
• 2024

3D Cross-Modal Retrieval (3DCMR) is a task that retrieves 3D objects regardless of modalities, such as images, meshes, and point clouds. One of the most prominent methods used for 3DCMR is the Cross-Modal Center Loss Function (CLF) which applies the conventional center loss strategy for 3D cross-modal search and retrieval. Since CLF is based on center loss, the center features in CLF are also susceptible to subtle changes in hyperparameters and external inferences. For instance, performance degradation is observed when the batch size is too small. Furthermore, the Mean Squared Error (MSE) used in CLF is unable to adapt to changes in batch size and is vulnerable to data variations that occur during actual inference due to the use of simple Euclidean distance between multi-modal features. To address the problems that arise from small batch training, we propose a Noisy Center Loss (NCL) method to estimate the optimal center features. In addition, we apply the simple Siamese representation learning method (SimSiam) during optimal center feature estimation to compare projected features, making the proposed method robust to changes in batch size and variations in data. As a result, the proposed approach demonstrates improved performance in ModelNet40 dataset compared to the conventional methods.