• Journal of radiological science and technology
• /
• v.45 no.1
• /
• pp.41-47
• /
• 2022

The purpose of this study is to apply a deep learning model that can distinguish lung perfusion and lung ventilation images in nuclear medicine, and to evaluate the image classification ability. Image data pre-processing was performed in the following order: image matrix size adjustment, min-max normalization, image center position adjustment, train/validation/test data set classification, and data augmentation. The convolutional neural network(CNN) structures of VGG-16, ResNet-18, Inception-ResNet-v2, and SE-ResNeXt-101 were used. For classification model evaluation, performance evaluation index of classification model, class activation map(CAM), and statistical image evaluation method were applied. As for the performance evaluation index of the classification model, SE-ResNeXt-101 and Inception-ResNet-v2 showed the highest performance with the same results. As a result of CAM, cardiac and right lung regions were highly activated in lung perfusion, and upper lung and neck regions were highly activated in lung ventilation. Statistical image evaluation showed a meaningful difference between SE-ResNeXt-101 and Inception-ResNet-v2. As a result of the study, the applicability of the CNN model for lung scintigraphy classification was confirmed. In the future, it is expected that it will be used as basic data for research on new artificial intelligence models and will help stable image management in clinical practice.

• Steel and Composite Structures
• /
• v.52 no.2
• /
• pp.145-163
• /
• 2024

This paper presents six novel hybrid machine learning (ML) models that combine support vector machines (SVM), Decision Tree (DT), Random Forest (RF), Gradient Boosting (GB), extreme gradient boosting (XGB), and categorical gradient boosting (CGB) with the Harris Hawks Optimization (HHO) algorithm. These models, namely HHO-SVM, HHO-DT, HHO-RF, HHO-GB, HHO-XGB, and HHO-CGB, are designed to predict the ultimate strength of both rectangular and circular reinforced concrete (RC) columns. The prediction models are established using a comprehensive database consisting of 325 experimental data for rectangular columns and 172 experimental data for circular columns. The ML model hyperparameters are optimized through a combination of cross-validation technique and the HHO. The performance of the hybrid ML models is evaluated and compared using various metrics, ultimately identifying the HHO-CGB model as the top-performing model for predicting the ultimate shear strength of both rectangular and circular RC columns. The mean R-value and mean a20-index are relatively high, reaching 0.991 and 0.959, respectively, while the mean absolute error and root mean square error are low (10.302 kN and 27.954 kN, respectively). Another comparison is conducted with four existing formulas to further validate the efficiency of the proposed HHO-CGB model. The Shapely Additive Explanations method is applied to analyze the contribution of each variable to the output within the HHO-CGB model, providing insights into the local and global influence of variables. The analysis reveals that the depth of the column, length of the column, and axial loading exert the most significant influence on the ultimate shear strength of RC columns. A user-friendly graphical interface tool is then developed based on the HHO-CGB to facilitate practical and cost-effective usage.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.5
• /
• pp.1035-1040
• /
• 2024

The main objective of inspecting structures is to ensure the safety of all entities that utilize these structures as cracks in structures if not attended to could lead to serious calamities. With that objective in mind, artificial intelligence (AI) based technologies to assist human inspectors are needed especially for retaining walls in structures. In this paper, we predict the crack displacement of retaining walls using an Polynomial Regressive (PR) analysis model, as well as Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) deep learning models, and compare their performance. For the performance comparison, we apply multi-variable feature inputs, by utilizing temperature and rainfall data that may affect the crack displacement of the retaining wall. The training and inference data were collected through measuring sensors such as inclinometers, thermometers, and rain gauges. The results show that the multi-variable feature model had a MAE of 0.00186, 0.00450 and 0.00842, which outperformed the single variable feature model at 0.00393, 0.00556 and 0.00929 for the polynomial regression model, LSTM model and the GRU model respectively from the evaluation performed.

• Journal of Wetlands Research
• /
• v.26 no.2
• /
• pp.147-159
• /
• 2024

Lack of streamflow observations makes model calibration difficult and limits model performance improvement. Satellite-based remote sensing products offer a new alternative as they can be actively utilized to obtain hydrological data. Recently, several studies have shown that artificial intelligence-based solutions are more appropriate than traditional conceptual and physical models. In this study, a data-driven approach combining various recurrent neural networks and decision tree-based algorithms is proposed, and the utilization of satellite remote sensing information for AI training is investigated. The satellite imagery used in this study is from MODIS and SMAP. The proposed approach is validated using publicly available data from 25 watersheds. Inspired by the traditional regionalization approach, a strategy is adopted to learn one data-driven model by integrating data from all basins, and the potential of the proposed approach is evaluated by using a leave-one-out cross-validation regionalization setting to predict streamflow from different basins with one model. The GRU + Light GBM model was found to be a suitable model combination for target basins and showed good streamflow prediction performance in ungauged basins (The average model efficiency coefficient for predicting daily streamflow in 25 ungauged basins is 0.7187) except for the period when streamflow is very small. The influence of satellite remote sensing information was found to be up to 10%, with the additional application of satellite information having a greater impact on streamflow prediction during low or dry seasons than during wet or normal seasons.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.4
• /
• pp.669-677
• /
• 2020

As the number of Internet users exploded, attacks on the web increased. In addition, the attack patterns have been diversified to bypass existing defense techniques. Traditional web firewalls are difficult to detect attacks of unknown patterns.Therefore, the method of detecting abnormal behavior by artificial intelligence has been studied as an alternative. Specifically, attempts have been made to apply natural language processing techniques because the type of script or query being exploited consists of text. However, because there are many unknown words in scripts and queries, natural language processing requires a different approach. In this paper, we propose a new classification model which uses byte pair encoding (BPE) technology to learn the embedding vector, that is often used for web attack payloads, and uses an attention mechanism-based Bi-GRU neural network to extract a set of tokens that learn their order and importance. For major web attacks such as SQL injection, cross-site scripting, and command injection attacks, the accuracy of the proposed classification method is about 0.9990 and its accuracy outperforms the model suggested in the previous study.

• Journal of KIISE:Software and Applications
• /
• v.27 no.4
• /
• pp.412-421
• /
• 2000

In spite that the main contents of mathematical and scientific learning are understanding principles and their applications, most of existing educational softwares are based on rote learning, thus resulting in limited educational effects. In the artificial intelligence research, some progress has been made in developing automatic tutors based on proving and simulation, by adapting the techniques of knowledge representation, search and inference to the design of tutors. However, these tutors still fall short of being practical and the turor, even a prototype model, for learning problem solving is yet to come out. The geometry problem-solving tutor proposed by this research involves dynamic inference performed in parallel with learning. As an ontology for composing the problem space within a real-time setting, we have employed the notions of propositions, hypotheses and operators. Then we investigated the mechanism of interactive learning of problem solving in which the main target of inference involves the generation and the test of these components. Major accomplishment from this research is a practical model of a problem tutor embedded with a series of inference techniques for algebraic manipulation, which is indispensable in geometry problem solving but overlooked by previous research. The proposed model is expected to be applicable to the design of problem tutors in other scientific areas such as physics and electric circuitry.

• Journal of Broadcast Engineering
• /
• v.24 no.1
• /
• pp.77-86
• /
• 2019

This paper proposes using machine-learning technology to analyze and classify historical collected documents based on them. Data is collected based on keywords associated with a specific domain and the non-conceptuals such as special characters are removed. Then, tag each word of the document collected using a Korean-language morpheme analyzer with its nouns, verbs, and sentences. Embedded documents using Doc2Vec model that converts documents into vectors. Measure the similarity between documents through the embedded model and learn the document classifier using the machine running algorithm. The highest performance support vector machine measured 0.83 of F1-score as a result of comparing the classification model learned.

• Phonetics and Speech Sciences
• /
• v.13 no.3
• /
• pp.71-78
• /
• 2021

Most neural-network-based speech synthesis models utilize neural vocoders to convert mel-scaled spectrograms into high-quality, human-like voices. However, neural vocoders combined with mel-scaled spectrogram prediction models demand considerable computer memory and time during the training phase and are subject to slow inference speeds in an environment where GPU is not used. This problem does not arise in linear spectrogram prediction models, as they do not use neural vocoders, but these models suffer from low voice quality. As a solution, this paper proposes a Tacotron 2 and Transformer-based linear spectrogram prediction model that produces high-quality speech and does not use neural vocoders. Experiments suggest that this model can serve as the foundation of a high-quality text-to-speech model with fast inference speed.

• The Journal of the Korea Contents Association
• /
• v.21 no.8
• /
• pp.1-9
• /
• 2021

Deep learning in computer vision has made accelerated improvement over a short period but large-scale learning data and computing power are still essential that required time-consuming trial and error tasks are involved to derive an optimal network model. In this study, we propose a similar image classification performance improvement method based on CR (Confusion Rate) that considers only the characteristics of the data itself regardless of network optimization or data reinforcement. The proposed method is a technique that improves the performance of the deep learning model by calculating the CRs for images in a dataset with similar characteristics and reflecting it in the weight of the Loss Function. Also, the CR-based recognition method is advantageous for image identification with high similarity because it enables image recognition in consideration of similarity between classes. As a result of applying the proposed method to the Resnet18 model, it showed a performance improvement of 0.22% in HanDB and 3.38% in Animal-10N. The proposed method is expected to be the basis for artificial intelligence research using noisy labeled data accompanying large-scale learning data.

• The Journal of the Korea Contents Association
• /
• v.21 no.11
• /
• pp.785-792
• /
• 2021

Since the Fourth Industrial Revolution, various changes have occurred in society as a whole due to advance in technologies such as artificial intelligence and big data. The amount of data that can be collect in the process of applying important technologies tends to increase rapidly. Especially in academia, existing generated literature data is analyzed in order to grasp research trends, and analysis of these literature organizes the research flow and organizes some research methodologies and themes, or by grasping the subjects that are currently being talked about in academia, we are making a lot of contributions to setting the direction of future research. However, it is difficult to access whether data collection is necessary for the analysis of document data without the expertise of ordinary programs. In this paper, propose a text mining-based topic modeling Web application model. Even if you lack specialized knowledge about data analysis methods through the proposed model, you can perform various tasks such as collecting, storing, and text-analyzing research papers, and researchers can analyze previous research and research trends. It is expect that the time and effort required for data analysis can be reduce order to understand.