• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권1호
• /
• pp.305-326
• /
• 2019

Discriminative correlation filter (DCF) based tracking algorithms have recently shown impressive performance on benchmark datasets. However, amount of recent researches are vulnerable to heavy occlusions, irregular deformations and so on. In this paper, we intend to solve these problems and handle the contradiction between accuracy and real-time in the framework of tracking-by-detection. Firstly, we propose an innovative strategy to combine the template and color-based models instead of a simple linear superposition and rely on the strengths of both to promote the accuracy. Secondly, to enhance the discriminative power of the learned template model, the spatial regularization is introduced in the learning stage to penalize the objective boundary information corresponding to features in the background. Thirdly, we utilize a discriminative multi-scale estimate method to solve the problem of scale variations. Finally, we research strategies to limit the computational complexity of our tracker. Abundant experiments demonstrate that our tracker performs superiorly against several advanced algorithms on both the OTB2013 and OTB2015 datasets while maintaining the high frame rates.

• Structural Engineering and Mechanics
• /
• 제69권1호
• /
• pp.21-31
• /
• 2019

Optimizing composite structures to exploit their maximum potential is a realistic application with promising returns. In this research, simultaneous maximization of the fundamental frequency and frequency separation between the first two modes by optimizing the fiber angles is considered. A high-fidelity design optimization methodology is developed by combining the high-accuracy of finite element method with iterative improvement capability of metaheuristic algorithms. Three powerful nature-inspired optimization algorithms viz. a genetic algorithm (GA), a particle swarm optimization (PSO) variant and a cuckoo search (CS) variant are used. Advanced memetic features are incorporated in the PSO and CS to form their respective variants-RPSOLC (repulsive particle swarm optimization with local search and chaotic perturbation) and CHP (co-evolutionary host-parasite). A comprehensive set of benchmark solutions on several new problems are reported. Statistical tests and comprehensive assessment of the predicted results show CHP comprehensively outperforms RPSOLC and GA, while RPSOLC has a little superiority over GA. Extensive simulations show that the on repeated trials of the same experiment, CHP has very low variability. About 50% fewer variations are seen in RPSOLC as compared to GA on repeated trials.

• Genomics & Informatics
• /
• 제17권2호
• /
• pp.17.1-17.3
• /
• 2019

Text mining has become an important research method in biology, with its original purpose to extract biological entities, such as genes, proteins and phenotypic traits, to extend knowledge from scientific papers. However, few thorough studies on text mining and application development, for plant molecular biology data, have been performed, especially for rice, resulting in a lack of datasets available to solve named-entity recognition tasks for this species. Since there are rare benchmarks available for rice, we faced various difficulties in exploiting advanced machine learning methods for accurate analysis of the rice literature. To evaluate several approaches to automatically extract information from gene/protein entities, we built a new dataset for rice as a benchmark. This dataset is composed of a set of titles and abstracts, extracted from scientific papers focusing on the rice species, and is downloaded from PubMed. During the 5th Biomedical Linked Annotation Hackathon, a portion of the dataset was uploaded to PubAnnotation for sharing. Our ultimate goal is to offer a shared task of rice gene/protein name recognition through the BioNLP Open Shared Tasks framework using the dataset, to facilitate an open comparison and evaluation of different approaches to the task.

• International journal of advanced smart convergence
• /
• 제10권3호
• /
• pp.41-50
• /
• 2021

The purpose of this study is to analyze various energy platforms using data in the US and to suggest directions and implications. Some of the leading energy platforms are selected and analyzed based on the opening platform model. We focus on the case analysis of the US utility companies. In case of the horizontal open platform, Green Button sponsor's 'Connect My Data (CMD)' driven by the government invites the utility companies to jointly develop the sponsor's data solution. In case of the vertical open platform, the certification program 'Share My Data (SMD)' allows backward compatibility, because the technical improvement is minimal. The utility companies benchmark Amazon's three-sided market mediation and prefer platform and category exclusivity. For the former, they have data analytics companies like Enervee, Opower and for the latter, they have electronics manufactures and energy service providers (ESPs) like Distributed Energy Resources (DERs). Based on this US case study, we suggest the energy platforms to open their platform for renewable energy supply, energy conservation, high-efficiency products, and residential DER dissemination. To successfully implement the government's energy transition policy, the US platforms should be benchmarked as a business model. Especially, it is needed for them to coordinate a platform ecosystem. To ensure trust in the products and services offered on the marketplace platform, platform's certification program is helpful.

• International journal of advanced smart convergence
• /
• 제9권4호
• /
• pp.16-25
• /
• 2020

The operation of a wastewater treatment plant (WWTP) is a complex task which requires to consider several aspects: adapting to always changing influent composition and volume, ensuring treated effluents quality complies with local regulations, ensuring dissolved oxygen levels in biological reaction tanks are sufficient to avoid anoxic conditions etc. all of it while minimizing usage of chemicals and power consumption. The traditional way of managing WWTPs consists in having employees on the field measure various parameters and make decisions based on their judgment and experience which holds various concerns such as the low frequency of data, errors in measurement and difficulty to analyze historical data to propose optimal solutions. In the case of activated sludge WWTPs, parts of the treatment process can be automated and controlled in order to satisfy various control objectives. The models developed by the International Water Association (IWA) have been extensively used worldwide in order to design and assess the performance of various control strategies. In this work, we propose to review most recent WWTP automation initiatives around the world and identify most currently used control parameters and control architectures. We then suggest a framework to select WWTP model, control parameters and control scheme in order to develop and benchmark control strategies for WWTP automation.

• Nuclear Engineering and Technology
• /
• 제54권9호
• /
• pp.3429-3439
• /
• 2022

The MCS code is a computer code developed by the Ulsan National Institute of Science and Technology (UNIST) for simulation and calculation of nuclear reactor systems based on the Monte Carlo method. The code is currently used to solve two main types of reactor physics problems, namely, criticality problems and radiation shielding problems. In this paper, the radiation shielding capability of the MCS code is validated by simulating some selected SINBAD (Shielding Integral Benchmark Archive and Database) experiments. The whole validation was performed in two ways. Firstly, the functionality and computational rationality of the MCS code was verified by comparing the simulation results with those of MCNP code. Secondly, the validity and computational accuracy of the MCS code was confirmed by comparing the simulation results with the experimental results of SINBAD. The simulation results of the MCS code are highly consistent with the those of the MCNP code, and they are within the 2σ error bound of the experiment results. It shows that the calculation results of the MCS code are reliable when simulating the radiation shielding problems.

• International Journal of Computer Science & Network Security
• /
• 제21권7호
• /
• pp.182-190
• /
• 2021

Deep learning is an advanced technology for large-scale data analysis, with numerous promising cases like image processing, object detection and significantly more. It becomes customarily to use transfer learning and fine-tune a pre-trained CNN model for most image recognition tasks. Having people taking photos and tag themselves provides a valuable resource of in-data. However, these tags and labels might be noisy as people who annotate these images might not be experts. This paper aims to explore the impact of noisy labels on fine-tuning pre-trained CNN models. Such effect is measured on a food recognition task using Food101 as a benchmark. Four pre-trained CNN models are included in this study: InceptionV3, VGG19, MobileNetV2 and DenseNet121. Symmetric label noise will be added with different ratios. In all cases, models based on DenseNet121 outperformed the other models. When noisy labels were introduced to the data, the performance of all models degraded almost linearly with the amount of added noise.

• 유통과학연구
• /
• 제19권6호
• /
• pp.41-50
• /
• 2021

Purpose: The importance of entrepreneurship education is increasing as interest in entrepreneurship is increasing around the world. In addition to the US and Europe, which operate advanced entrepreneurship education programs, Korea is also investing a lot of government resources for university centered entrepreneurship education. On the other hand, developing countries, which have high interest in Korea's economic development, are also trying to benchmark this Korean university entrepreneurship education. Research design, data and methodology: This study systematizes experiences such as training on entrepreneurship, science park management, and policy consulting for science, technology parks, and universities in developing countries. Through this, the needs of the relevant countries are analyzed based on the results of previous research, related theories, and policies. Results: As a result of the analysis, four key elements were derived for the establishment of entrepreneurship education and entrepreneurship ecosystem in developing countries. In addition, the details that these elements can be used in the university entrepreneurship ecosystem are presented in the form of tasks in stages. Conclusions: This study presents factors, including entrepreneurship-based leadership and human resources, structure and program, domestic and international network, and budget as a plan for revitalizing entrepreneurship education in developing countries.

• 국제학술발표논문집
• /
• The 9th International Conference on Construction Engineering and Project Management
• /
• pp.814-822
• /
• 2022

The deployment of sensors for Structural Health Monitoring requires a complicated network arrangement, ground truthing, and calibration for validating sensor performance periodically. Any conventional sensor on a structural element is also subjected to static and dynamic vertical loadings in conjunction with other environmental factors, such as brightness, noise, temperature, and humidity. A structural model with strain gauges was built and tested to get realistic sensory information. This paper investigates different deep learning architectures and algorithms, including unsupervised, autoencoder, and supervised methods, to benchmark blind drift calibration methods using deep learning. It involves a fully connected neural network (FCNN), a long short-term memory (LSTM), and a gated recurrent unit (GRU) to address the blind drift calibration problem (i.e., performing calibrations of installed sensors when ground truth is not available). The results show that the supervised methods perform much better than unsupervised methods, such as an autoencoder, when ground truths are available. Furthermore, taking advantage of time-series information, the GRU model generates the most precise predictions to remove the drift overall.

• Steel and Composite Structures
• /
• 제47권1호
• /
• pp.1-17
• /
• 2023

This paper deals with the free vibration behavior of rotating composite beams reinforced with carbon nanotubes (CNTs) under uniform thermal loads. The temperature-dependent beam material is assumed to be a mixture of single-walled carbon nanotubes (SWCNTs) in an isotropic matrix and five different functionally graded (FG) distributions of CNTs are considered according to the variation along the thickness, namely the UD-uniform, FG-O, FG-V, FG-Λ and FG-X distributions where FG-V and FG-Λ are unsymmetrical patterns. Considering the Timoshenko beam theory (TBT), a new finite element formulation of functionally graded carbon nanotube reinforced composite (FGCNTRC) beam is created for the first time. And the effects of several essential parameters including rotational speed, hub radius, effective material properties, slenderness ratio, boundary conditions, thermal force and moments due to temperature variation are considered in the formulation. By implementing different boundary conditions, some new results of both symmetric and non-symmetrical distribution patterns are presented in tables and figures to be used as benchmark for further validation. In addition, as an alternative advanced composite application for rotating systems exposed to thermal load, the positive effects of CNT addition in improving the dynamic performance of the system have been observed and the results are presented in several tables and figures.