• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.6
• /
• pp.285-292
• /
• 2023

With the wide spread of smart farms and the advancements in IoT technology, it is easy to obtain additional data in addition to crop images. Consequently, deep learning-based crop disease diagnosis research utilizing multimodal data has become important. This study proposes a crop disease diagnosis method using multimodal supervised contrastive learning by expanding upon the multimodal self-supervised learning. RandAugment method was used to augment crop image and time series of environment data. These augmented data passed through encoder and projection head for each modality, yielding low-dimensional features. Subsequently, the proposed multimodal supervised contrastive loss helped features from the same class get closer while pushing apart those from different classes. Following this, the pretrained model was fine-tuned for crop disease diagnosis. The visualization of t-SNE result and comparative assessments of crop disease diagnosis performance substantiate that the proposed method has superior performance than multimodal self-supervised learning.

• Korean Journal of Environmental Biology
• /
• v.41 no.4
• /
• pp.608-615
• /
• 2023

The morphological features of germling cells were examined to identify an unspecified resting cyst (described as Cochlodinium cf. polykrikoides-like resting cyst) in the Korean coastal area. LSU rRNA gene sequences were also obtained from a strain established from the germling cells. The resting cysts isolated from Korean coastal sediment were characterized as being brown in color, having a large dark-red body, and fibrous lobed ornaments. The germling cells were ellipsoidal with an irregular outline and had an open comma-shaped ASC (apical structure complex), a wide and deep cingulum, and a deep sulcus. These morphological features were consistent with those of previously described harmful dinoflagellate Pseudocochlodinium profundisulcus. The molecular phylogeny revealed that the germling cells and P. profundisulcus were conspecific. Based on these morphological and phylogenetic data, this study documents the occurrence of P. profundisulcus in a Korean coastal area for the first time.

• Journal of Auto-vehicle Safety Association
• /
• v.16 no.3
• /
• pp.18-24
• /
• 2024

Recent advancements in vehicle driving data analysis have attracted significant attention as a key research area aimed at optimizing driving behavior and enhancing vehicle performance. This study aims to collect operational data from drivers, including steering angle, longitudinal acceleration, brake usage, and wheel speed, and to cluster driving segments based on their characteristics using deep learning techniques. To achieve this, we employed a methodology that integrates the LSTM-Autoencoder model, which captures essential features of time-series data through compression and reconstruction, with an Attention mechanism. The driving segments were clustered and visually represented with distinct colors for clarity. Experimental results from the proposed approach demonstrated an over 96% accuracy in aligning input and output values, facilitating the clustering of driving segments based on their distinctive features. These findings provide valuable insights for improving driver behavior and optimizing vehicle assistance systems, potentially contributing to significant advancements in this field of research.

• The Bulletin of The Korean Astronomical Society
• /
• v.41 no.2
• /
• pp.31.2-31.2
• /
• 2016

We present our study of the correlation between the UV luminosity and the merging activities of nearby galaxies (d < 300 Mpc). Our study uses ~600 UV-selected galaxies with deep optical imaging data, where the UV selection is made using the GALEX Atlas of Galaxies (Gil de Paz et al. 2007) and the updated UV catalog of nearby galaxies (Yu Bai et al. 2015). Deep optical images allow us to classify merger features using visual inspection, and we also estimate unobscured SFR using UV continuum luminosity. The fraction of galaxies with merger features in each UV luminosity bins are obtained to see if how the fraction of galaxies with merging features changes as a function of UV luminosity, Finally, we will show, above what UV luminosity (or SFR), the merging mechanism becomes an important process in enhancing star formation of galaxies.

• International journal of advanced smart convergence
• /
• v.8 no.2
• /
• pp.132-139
• /
• 2019

We address the problem about forecasting the direction of stock price movement in the Korea market. Recently, the deep neural network is popularly applied in this area of research. In deep neural network systems, proper parameter initialization reduces training time and improves the performance of the model. Therefore, in our study, we propose a novel parameter initialization technique and apply this technique for the stock price movement prediction model. Specifically, we design a framework which consists of two models: a base model and a main prediction model. The base model constructed with LSTM is trained by using the large data which is generated by a large amount of the stock data to achieve optimal parameters. The main prediction model with the same architecture as the base model uses the optimal parameter initialization. Thus, the main prediction model is trained by only using the data of the given stock. Moreover, the stock price movements can be affected by other related information in the stock market. For this reason, we conducted our research with two types of inputs. The first type is the stock features, and the second type is a combination of the stock features and the Korea Composite Stock Price Index (KOSPI) features. Empirical results conducted on the top five stocks in the KOSPI list in terms of market capitalization indicate that our approaches achieve better predictive accuracy and F1-score comparing to other baseline models.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.6
• /
• pp.2018-2043
• /
• 2022

Nowadays, COVID-19 infections are influencing our daily lives which have spread globally. The major symptoms' of COVID-19 are dry cough, sore throat, and fever which in turn to critical complications like multi organs failure, acute respiratory distress syndrome, etc. Therefore, to hinder the spread of COVID-19, a Computerized Doughty Predictor Framework (CDPF) is developed to yield benefits in monitoring the progression of disease from Chest CT images which will reduce the mortality rates significantly. The proposed framework CDPF employs Convolutional Neural Network (CNN) as a feature extractor to extract the features from CT images. Subsequently, the extracted features are fed into the Adaptive Dragonfly Algorithm (ADA) to extract the most significant features which will smoothly drive the diagnosing of the COVID and Non-COVID cases with the support of Doughty Learners (DL). This paper uses the publicly available SARS-CoV-2 and Github COVID CT dataset which contains 2482 and 812 CT images with two class labels COVID⁺ and COVI^-. The performance of CDPF is evaluated against existing state of art approaches, which shows the superiority of CDPF with the diagnosis accuracy of about 99.76%.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.5
• /
• pp.1597-1610
• /
• 2022

At present, the main method of high-speed train chassis detection is using computer vision technology to extract keypoints from two related chassis images firstly, then matching these keypoints to find the pixel-level correspondence between these two images, finally, detection and other steps are performed. The quality and accuracy of image matching are very important for subsequent defect detection. Current traditional matching methods are difficult to meet the actual requirements for the generalization of complex scenes such as weather, illumination, and seasonal changes. Therefore, it is of great significance to study the high-speed train image matching method based on deep learning. This paper establishes a high-speed train chassis image matching dataset, including random perspective changes and optical distortion, to simulate the changes in the actual working environment of the high-speed rail system as much as possible. This work designs a convolutional neural network to intensively extract keypoints, so as to alleviate the problems of current methods. With multi-level features, on the one hand, the network restores low-level details, thereby improving the localization accuracy of keypoints, on the other hand, the network can generate robust keypoint descriptors. Detailed experiments show the huge improvement of the proposed network over traditional methods.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.881-891
• /
• 2023

Recently, various studies on malicious URL detection using artificial intelligence have been conducted, and most of the research have shown great detection performance. However, not only does classical machine learning require a process of analyzing features, but the detection performance of a trained model also depends on the data analyst's ability. In this paper, we propose a DL-ML Fusion Hybrid Model for malicious web site URL detection based on URL lexical features. the propose model combines the automatic feature extraction layer of deep learning and classical machine learning to improve the feature engineering issue. 60,000 malicious and normal URLs were collected for the experiment and the results showed 23.98%p performance improvement in maximum. In addition, it was possible to train a model in an efficient way with the automation of feature engineering.

• International Journal of Computer Science & Network Security
• /
• v.23 no.12
• /
• pp.213-219
• /
• 2023

In recent years, popularity of deep learning (DL) is increased due to its ability to extract features from Hyperspectral images. A lack of discrimination power in the features produced by traditional machine learning algorithms has resulted in poor classification results. It's also a study topic to find out how to get excellent classification results with limited samples without getting overfitting issues in hyperspectral images (HSIs). These issues can be addressed by utilising a new learning network structure developed in this study.EfficientNet-B4-Based Convolutional network (EN-B4), which is why it is critical to maintain a constant ratio between the dimensions of network resolution, width, and depth in order to achieve a balance. The weight of the proposed model is optimized by Search and Rescue Operations (SRO), which is inspired by the explorations carried out by humans during search and rescue processes. Tests were conducted on two datasets to verify the efficacy of EN-B4, with Indian Pines (IP) and the University of Pavia (UP) dataset. Experiments show that EN-B4 outperforms other state-of-the-art approaches in terms of classification accuracy.

• Journal of Information Processing Systems
• /
• v.20 no.4
• /
• pp.535-549
• /
• 2024

Facial expressions (FEs) serve as fundamental components for human emotion assessment and human-computer interaction. Traditional convolutional neural networks tend to overlook valuable information during the FE feature extraction, resulting in suboptimal recognition rates. To address this problem, we propose a deep learning framework that incorporates hierarchical feature fusion, contextual data, and an attention mechanism for precise FE recognition. In our approach, we leveraged an enhanced VGGNet16 as the backbone network and introduced an improved group convolutional channel attention (GCCA) module in each block to emphasize the crucial expression features. A partial decoder was added at the end of the backbone network to facilitate the fusion of multilevel features for a comprehensive feature map. A reverse attention mechanism guides the model to refine details layer-by-layer while introducing contextual information and extracting richer expression features. To enhance feature distinguishability, we employed islanding loss in combination with softmax loss, creating a joint loss function. Using two open datasets, our experimental results demonstrated the effectiveness of our framework. Our framework achieved an average accuracy rate of 74.08% on the FER2013 dataset and 98.66% on the CK+ dataset, outperforming advanced methods in both recognition accuracy and stability.