• Journal of Information Processing Systems
• /
• v.15 no.5
• /
• pp.1179-1191
• /
• 2019

We design an ingenious view-pooling method named learning-based multiple pooling fusion (LMPF), and apply it to multi-view convolutional neural network (MVCNN) for 3D model classification or retrieval. By this means, multi-view feature maps projected from a 3D model can be compiled as a simple and effective feature descriptor. The LMPF method fuses the max pooling method and the mean pooling method by learning a set of optimal weights. Compared with the hand-crafted approaches such as max pooling and mean pooling, the LMPF method can decrease the information loss effectively because of its "learning" ability. Experiments on ModelNet40 dataset and McGill dataset are presented and the results verify that LMPF can outperform those previous methods to a great extent.

• ETRI Journal
• /
• v.42 no.6
• /
• pp.886-898
• /
• 2020

Unusual data patterns or outliers can be generated because of human errors, incorrect measurements, or malicious activities. Detecting outliers is a difficult task that requires complex ensembles. An ideal outlier detection ensemble should consider the strengths of individual base detectors while carefully combining their outputs to create a strong overall ensemble and achieve unbiased accuracy with minimal variance. Selecting and combining the outputs of dissimilar base learners is a challenging task. This paper proposes a model that utilizes heterogeneous base learners. It adaptively boosts the outcomes of preceding learners in the first phase by assigning weights and identifying high-performing learners based on their local domains, and then carefully fuses their outcomes in the second phase to improve overall accuracy. Experimental results from 10 benchmark datasets are used to train and test the proposed model. To investigate its accuracy in terms of separating outliers from inliers, the proposed model is tested and evaluated using accuracy metrics. The analyzed data are presented as crosstabs and percentages, followed by a descriptive method for synthesis and interpretation.

• International journal of steel structures
• /
• v.18 no.5
• /
• pp.1741-1753
• /
• 2018

Metallic dampers are sacrificial devices (fuses) that dissipate significant energy during earthquakes while protecting other parts of structures from possible damage. In addition to numerous implementation opportunities of other base isolation systems, U-shaped dampers (UD) are one of the widely investigated and used devices in practice especially in Japan. The present study focuses on enhancing seismic performance of these types of dampers by changing their geometric properties. UDs with perforated (i.e. with holes) and/or nonparallel arms are developed for this purpose. For a better comparison, the criterion of equal material volume (or mass) has been utilized. Three dimensional finite element models of the new type of UDs are formed and investigated numerically under selected displacement histories. Based on the obtained hysteretic curves; dissipated energy intensities, effective stiffness ratios, reaction forces, effective damping ratios are evaluated in this parametric study. It is found that both damper types have merits in use of seismic applications and that the selection of the damper configuration is dependent on the design specific issues.

• Structural Engineering and Mechanics
• /
• v.81 no.2
• /
• pp.147-161
• /
• 2022

The Brace-Type Shear Fuse (BSF) device is a newly proposed steel damper with excellent cumulative ductility and stable energy dissipation. In consideration of the current situation where there are not many alternatives for transversal seismic devices used in long-span three-tower self-anchored bridges (TSSBs), this paper implements improved BSFs into the world's longest TSSB, named Jinan Fenghuang Yellow River Bridge. The new details of the BSF are developed for the TSSB, and the force-displacement hysteretic curves of the BSFs are obtained using finite element (FE) simulations. A three-dimensional refined finite element model for the research TSSB was established in SAP2000, and the effects of BSFs on dynamic characteristics and seismic response of the TSSB under different site conditions were investigated by the numerical simulation method. The results show that remarkable controlling effects of BSFs on seismic response of TSSBs under different site conditions were obtained. Compared with the case without BSFs, the TSSB installed with BSFs has mitigation ratios of the tower top displacement, lateral girder displacement, tower bending moment and tower shear force exceeding 95%, 78%, 330% and 346%, respectively. Meanwhile, BSFs have a sufficient restoring force mechanism with a minor post-earthquake residual displacement. The proposed BSFs exhibit good application prospects in long-span TSSBs.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.354-357
• /
• 2021

Infants communicate their feelings and needs to the outside world through non-verbal methods such as crying and displaying diverse facial expressions. However, inexperienced parents tend to decode these non-verbal messages incorrectly and take inappropriate actions, which might affect the bonding they build with their babies and the cognitive development of the newborns. In this paper, we propose an aural-visual two-stream based infant cry recognition system to help parents comprehend the feelings and needs of crying babies. The proposed system first extracts the features from the pre-processed audio and video data by using the VGGish model and 3D-CNN model respectively, fuses the extracted features using a fully connected layer, and finally applies a SoftMax function to classify the fused features and recognize the corresponding type of cry. The experimental results show that the proposed system classification exceeds 0.92 in F1-score, which is 0.08 and 0.10 higher than the single-stream aural model and single-stream visual model.

• Steel and Composite Structures
• /
• v.46 no.3
• /
• pp.417-433
• /
• 2023

The maximum ductility and cumulative ductility of connection joints of Buckling-Restrained Braced Frames (BRBF) are critical to the structural overall performance, which should be matched with the BRB ductility. The two-story and one-span BRBF with a one-third scale was tested under cyclic quasi-static loading, and the top-flange beam splice (TFBS) rotational connections were proposed and adopted in BRBF. The deformation capacity of TFBS connections was observed during the test, and the relationship between structural global ductility and local connection ductility was studied. The rotational capacity of the beam-column connections and the stability performance of the BRBs are highly relevant to the structural overall performance. The hysteretic curves of BRBF are stable and full under large displacement demand imposed up to 2% story drift, and energy is dissipated as the large plastic deformation developed in the structural components. The BRBs acted as fuses and yielded first, and the cumulative plastic ductility (CPD) of BRBs is 972.6 of the second floor and 439.7 of the first floor, indicating the excellent energy dissipation capacity of BRBs. Structural members with good local ductility ensure the large global ductility of BRBF. The ductile capacity and hysteretic behavior of BRBF with TFBS connections were compared with those of BRBF with Reduced Beam Section (RBS) connections in terms of the experimental results.

• Journal of Distribution Science
• /
• v.21 no.12
• /
• pp.103-111
• /
• 2023

Purpose: This study explores the impact of live commerce on consumer purchase intention in the Chinese market. Live commerce, a novel marketing strategy that fuses e-commerce with live streaming, has gained significant popularity in China. The study aims to pinpoint main ways that live commerce increases customers' propensity to make purchases. Research design, data, and methodology: The PRISMA technique was adhered to in the systematic cycle of finding, screening, and selecting publications. Predefined keywords like "live commerce," "live-stream shopping," "consumer purchase intention," and "China" produced a large number of possibly pertinent research in the first search. Results: This study has delved into the profound effects of live commerce on consumer purchase intention, with a specific focus on the dynamic Chinese market. This study has identified four primary outcomes that support the revolutionary potential of live commerce: increased product engagement, significant social influences, dynamic pricing tactics, and the establishment of trust and credibility. Conclusions: Finally, the study provides live commerce practitioners with valuable insights, advising them on maximizing these impacts to increase consumer buy intention. In summary, this study offers helpful information to companies and marketers who want to take advantage of live commerce as a powerful tool for connecting with Chinese customers and increasing sales.

• Journal of Information Processing Systems
• /
• v.19 no.5
• /
• pp.673-687
• /
• 2023

Convolutional neural network-based deep learning technology is the most commonly used in image identification, but it requires large-scale data for training. Therefore, application in specific fields in which data acquisition is limited, such as in the military, may be challenging. In particular, the identification of ground weapon systems is a very important mission, and high identification accuracy is required. Accordingly, various studies have been conducted to achieve high performance using small-scale data. Among them, the ensemble method, which achieves excellent performance through the prediction average of the pre-trained models, is the most representative method; however, it requires considerable time and effort to find the optimal combination of ensemble models. In addition, there is a performance limitation in the prediction results obtained by using an ensemble method. Furthermore, it is difficult to obtain the ensemble effect using models with imbalanced classification accuracies. In this paper, we propose a transfer learning-based feature fusion technique for heterogeneous models that extracts and fuses features of pre-trained heterogeneous models and finally, fine-tunes hyperparameters of the fully connected layer to improve the classification accuracy. The experimental results of this study indicate that it is possible to overcome the limitations of the existing ensemble methods by improving the classification accuracy through feature fusion between heterogeneous models based on transfer learning.

• The Journal of the Korea Contents Association
• /
• v.20 no.4
• /
• pp.49-61
• /
• 2020

Recently, the convergence of genre-mixed collaboration and technology-mixed collaboration in performance art strategically fuses different goals and different properties, satisfying anticipated demand, and developing into experimental forms that bring out convergence contents of new value. To supplement problems of convergence attempts and heighten levels of completion, the effective collaborative processes of media experts must be studied and improved. This thesis attempts to study effective collaborative production plans of convergence performance that correspond to the demands of the times through multimedia convergence performance prototypes(Live performance play + Pre-made digital animation). It categorizes performance production processes into pre-production and production and researches the effective collaborative production processes of convergence performances that utilize these two forms of media through work selection, production direction establishment, human resource constituents, production schedule plan establishment, visualization processes, and performance practices and rehearsals. Continuous research must be conducted based on convergence performance contents planning, changes in production methods, and an understanding of distinct characteristics among convergence contents for the industrial development systemization and vitalization of convergence contents to be made possible.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.6
• /
• pp.3188-3207
• /
• 2017

In this paper, we propose an image enhancement algorithm to improve license plate extraction rate in various environments (Day Street, Night Street, Underground parking lot, etc.). The proposed algorithm is composed of image enhancement algorithm and license plate extraction algorithm. The image enhancement method can improve an image quality of the degraded image, which utilizes a histogram information and overall gray level distribution of an image. The proposed algorithm employs an interpolated probability distribution value (PDV) in order to control a sudden change in image brightness. Probability distribution value can be calculated using cumulative distribution function (CDF) and probability density function (PDF) of the captured image, whose values are achieved by brightness distribution of the captured image. Also, by adjusting the image enhancement factor of each part region based on image pixel information, it provides a function that can adjust the gradation of the image in more details. This processed gray image is converted into a binary image, which fuses narrow breaks and long thin gulfs, eliminates small holes, and fills gaps in the contour by using morphology operations. Then license plate region is detected based on aspect ratio and license plate size of the bound box drawn on connected license plate areas. The images have been captured by using a video camera or a personal image recorder installed in front of the cars. The captured images have included several license plates on multilane roads. Simulation has been executed using OpenCV and MATLAB. The results show that the extraction success rate is more improved than the conventional algorithms.