• Journal of Korean Society of Disaster and Security
• /
• v.16 no.4
• /
• pp.129-136
• /
• 2023

A special offshore bridge with a high pylon has special structural features.Special offshore bridges have inspection blind spots that are difficult to visually inspect. To solve this problem, safety inspection methods using drones are being studied. In this study, image data of the pylon of a special offshore bridge was acquired using a drone. In addition, an artificial intelligence algorithm was developed to detect damage to the pylon surface. The AI algorithm utilized a deep learning network with different structures. The algorithm applied the stacking ensemble learning method to build a model that formed the ensemble and collect the results.

• Journal of Digital Policy
• /
• v.3 no.1
• /
• pp.27-32
• /
• 2024

In this paper, proposes a super resolution method that enhances the quality of results by refining texture features, contrasting each, and utilizing the results as weights. For the improvement of quality, a precise and clear restoration result in details such as boundary areas is crucial in super resolution, along with minimizing unnecessary artifacts like noise. The proposed method constructs a residual block structure with multiple paths and skip-connections for feature estimation in conventional Convolutional Neural Network (CNN)-based super resolution methods to enhance quality. Additional learning is performed for sharpened and blurred image results for further texture analysis. By contrasting each super resolution result and allocating weights through this process, the proposed method achieves improved quality in detailed and smoothed areas of the image. The experimental results of the proposed method, evaluated using the PSNR and SSIM values as quality metrics, show higher results compared to existing algorithms, confirming the enhancement in quality.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.34 no.1
• /
• pp.103-114
• /
• 2024

In a situation where drone-related crimes continue to rise, research in drone forensics becomes crucial for preventing and responding to incidents involving drones. Conducting forensic analysis on flight record files stored internally is essential for investigating illegal activities. However, analyzing flight record files generated through the exclusive DUML protocol requires a deep understanding of the protocol's structure and characteristics. Additionally, a forensic analysis tool capable of handling cryptographic payloads and analyzing various drone models is imperative. Therefore, this study presents the methods and characteristics of flight record files generated by drones. It also explains the structure of the flight record file and the features of the DUML packet. Ultimately, we conduct forensic analysis based on the presented structure of the DUML packet and propose an extension forensic analysis system that operates more universally than existing tools, performing expanded syntactic analysis.

• Journal of the Korea Computer Graphics Society
• /
• v.30 no.3
• /
• pp.109-123
• /
• 2024

Research into vision-based end-to-end autonomous driving systems utilizing deep learning and reinforcement learning has been steadily increasing. These systems typically encode continuous and high-dimensional vehicle states, such as location, velocity, orientation, and sensor data, into latent features, which are then decoded into a vehicular control policy. The complexity of urban driving environments necessitates the use of state representation learning through networks like Variational Autoencoders (VAEs) or Convolutional Neural Networks (CNNs). This paper analyzes the impact of different image state encoding methods on reinforcement learning performance in autonomous driving. Experiments were conducted in the CARLA simulator using RGB images and semantically segmented images captured by the vehicle's front camera. These images were encoded using VAE and Vision Transformer (ViT) networks. The study examines how these networks influence the agents' learning outcomes and experimentally demonstrates the role of each state representation technique in enhancing the learning efficiency and decision- making capabilities of autonomous driving systems.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.12
• /
• pp.3218-3241
• /
• 2023

Financial fraud undermines the sustainable development of financial markets. Financial statements can be regarded as the key source of information to obtain the operating conditions of listed companies. Current research focuses more on mining financial digital data instead of looking into text data. However, text data can reveal emotional information, which is an important basis for detecting financial fraud. The audit opinion of the financial statement is especially the fair opinion of a certified public accountant on the quality of enterprise financial reports. Therefore, this research was carried out by using the data features of 4,153 listed companies' financial annual reports and audits of text opinions in the past six years, and the paper puts forward a financial fraud detection model integrating audit opinions. First, the financial data index database and audit opinion text database were built. Second, digitized audit opinions with deep learning Bert model was employed. Finally, both the extracted audit numerical characteristics and the financial numerical indicators were used as the training data of the LightGBM model. What is worth paying attention to is that the imbalanced distribution of sample labels is also one of the focuses of financial fraud research. To solve this problem, data enhancement and Focal Loss feature learning functions were used in data processing and model training respectively. The experimental results show that compared with the conventional financial fraud detection model, the performance of the proposed model is improved greatly, with Area Under the Curve (AUC) and Accuracy reaching 81.42% and 78.15%, respectively.

• Journal of Veterinary Clinics
• /
• v.41 no.4
• /
• pp.252-257
• /
• 2024

A 13-year-old castrated male mixed-breed dog presented for a health screening with a small nodule on the left hindlimb, which was revealed to be a mast cell tumor. A CT scan was performed to evaluate the mast cell tumor, and it incidentally revealed a well-marginated, fluid-attenuating, non-contrast enhanced mass in the deep ventral region of the left lower neck. Ultrasonography confirmed a well-capsulated mass with a cranio-ventrally protruding lesion extending into sternohyoid muscles. The outer layer of the mass had a hypoechoic thick wall and the internal parenchyma contained hyperechoic foci and multiple hyperechoic parallel lines. The mass was surgically resected. Histopathological examination confirmed a dermoid cyst characterized by cornifying squamous epithelial cells, keratinaceous debris, and hair shaft fragments. The dog showed no signs of recurrence or additional abnormalities three months post-surgery. This report highlights the importance of considering dermoid cysts in the differential diagnosis of neck masses on imaging examinations using CT or ultrasonography.

• Journal of IKEEE
• /
• v.28 no.3
• /
• pp.405-411
• /
• 2024

In this paper, we have proposed VmCUnet, a deep learning model designed to enhance image segmentation performance in skin lesion image. VmCUnet has combined Vm-UnetV2 with the CIM(Cross-Scale Interaction Module), and the features extracted from each layer of the encoder have been integrated through CIM to accurately recognize the boundaries of various patterns and objects. VmCUnet has performed image segmentation of skin lesions using ISIC-2017 and ISIC-2018 datasets and has outperformed Unet, TransUnet, SwinUnet, Vm-Unet, and Vm-UnetV2 on the performance metrics IoU and Dice Score. In future work, we will conduct additional experiments on different medical imaging datasets to validate the generalization performance of the VmCUnet model.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.37 no.2
• /
• pp.90-103
• /
• 2009

The poem 'Kyungjusipiyung(慶州十二詠)' written by Seo, Geo-jeong(徐居正) describes sentiments felt for the ruined historical and cultural landscape of Silla's capital city, Kyungju. It differs from the existing 'Eight Sceneries(八景)' as it conveys the strong metaphorical aesthetics of ruins as the episodes and figures are sung, as well as the myths and stories related to the representative holy places of the Silla culture: Gyelim(鷄林), Banwolseong(半月城), Najeong(蘿井), Oneung(五陵), Geumosan(金鰲山), the scenic beauty of deep placeness, Poseokjeong(鮑石亭), Mooncheon(蚊川), Cheomseongdae(瞻星臺), Boonhwangsa(芬皇寺), Youngmyosa(靈妙寺) and Grave of the General Kim Yu-Sin(金庾信墓). Compared with the former "Eight Sceneries" Poems, including Seo Geojeong's 'Kyungjusipiyung', there is a difference in the content of theme recitation, as well as in structure and form, especially with the deep impression of the classical features of the meanings and acts. The sequence of theme recitation seems to be composed of more than two visual corridors visited during trips that last longer than two days. The dominant emotions expresses in this poem, through written in the spring, are regret and sadness such as 'worn', 'broken and ruined', 'old and sad', without touching on the beauty of nature and the taste for life that is found in most of the Eight Sceneries Poems. Thus, the feelings of the reciter himself, Seo, Geo-jeong, about the described sceneries and their symbolism are more greatly emphasized than the beauty of form. The characteristic aspect of his experiences of ruins expressed from 'Kyungjusipiyung' is that the experiences were, first of all, qualitative of the aura conveyed; that is, the quality omnipresent throughout the culture of Silla as reflected in the twelve historical and cultural landscapes. In this poem, the cultural ruins of the invisible dimension such as the myths and legends are described by repetition, parallelism, juxtaposition, reflection and admiration from the antiphrases, as well as the civilized ruins of the visible dimension such as the various sceneries and features of Kyungju. This seems to be characteristic of the methods by which Seo, Geo-jeong appreciates 'Silla' in the poem 'Kyungjusipiyung'. Ruins as an Aesthetic Object imply the noble pride of Seo, Geo-jeong in identifying himself with the great nature of ruins. In 'Kyungjusipiyung', the images of the ruins of Silla and Kyungju are interspersed in spite of his positive recognition of 'the village of Kyungju' based on his records. However, though the concept of ruins has a pessimistic tone connoting the road of extinction and downfall, the aspect here seems to ambivalently contain the desire to recover and revive Kyungju through the Chosun Dynasty as adominant influence on the earlier Chosun's literary tide. The aesthetics of the scenery found in Seo, Geo-jeong's 'Kyungjusipiyung' contain the strongest of metaphor and symbolism by converting the experiences of the paradoxical ruins into the value of reflective experiences.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.163-177
• /
• 2019

As smartphones are getting widely used, human activity recognition (HAR) tasks for recognizing personal activities of smartphone users with multimodal data have been actively studied recently. The research area is expanding from the recognition of the simple body movement of an individual user to the recognition of low-level behavior and high-level behavior. However, HAR tasks for recognizing interaction behavior with other people, such as whether the user is accompanying or communicating with someone else, have gotten less attention so far. And previous research for recognizing interaction behavior has usually depended on audio, Bluetooth, and Wi-Fi sensors, which are vulnerable to privacy issues and require much time to collect enough data. Whereas physical sensors including accelerometer, magnetic field and gyroscope sensors are less vulnerable to privacy issues and can collect a large amount of data within a short time. In this paper, a method for detecting accompanying status based on deep learning model by only using multimodal physical sensor data, such as an accelerometer, magnetic field and gyroscope, was proposed. The accompanying status was defined as a redefinition of a part of the user interaction behavior, including whether the user is accompanying with an acquaintance at a close distance and the user is actively communicating with the acquaintance. A framework based on convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent networks for classifying accompanying and conversation was proposed. First, a data preprocessing method which consists of time synchronization of multimodal data from different physical sensors, data normalization and sequence data generation was introduced. We applied the nearest interpolation to synchronize the time of collected data from different sensors. Normalization was performed for each x, y, z axis value of the sensor data, and the sequence data was generated according to the sliding window method. Then, the sequence data became the input for CNN, where feature maps representing local dependencies of the original sequence are extracted. The CNN consisted of 3 convolutional layers and did not have a pooling layer to maintain the temporal information of the sequence data. Next, LSTM recurrent networks received the feature maps, learned long-term dependencies from them and extracted features. The LSTM recurrent networks consisted of two layers, each with 128 cells. Finally, the extracted features were used for classification by softmax classifier. The loss function of the model was cross entropy function and the weights of the model were randomly initialized on a normal distribution with an average of 0 and a standard deviation of 0.1. The model was trained using adaptive moment estimation (ADAM) optimization algorithm and the mini batch size was set to 128. We applied dropout to input values of the LSTM recurrent networks to prevent overfitting. The initial learning rate was set to 0.001, and it decreased exponentially by 0.99 at the end of each epoch training. An Android smartphone application was developed and released to collect data. We collected smartphone data for a total of 18 subjects. Using the data, the model classified accompanying and conversation by 98.74% and 98.83% accuracy each. Both the F1 score and accuracy of the model were higher than the F1 score and accuracy of the majority vote classifier, support vector machine, and deep recurrent neural network. In the future research, we will focus on more rigorous multimodal sensor data synchronization methods that minimize the time stamp differences. In addition, we will further study transfer learning method that enables transfer of trained models tailored to the training data to the evaluation data that follows a different distribution. It is expected that a model capable of exhibiting robust recognition performance against changes in data that is not considered in the model learning stage will be obtained.

• The Journal of Bigdata
• /
• v.4 no.2
• /
• pp.1-12
• /
• 2019

As Artificial Intelligence (AI) technology develops, it is applied to various fields such as image, voice, and text. AI has shown fine results in certain areas. Researchers have tried to predict the stock market by utilizing artificial intelligence as well. Predicting the stock market is known as one of the difficult problems since the stock market is affected by various factors such as economy and politics. In the field of AI, there are attempts to predict the ups and downs of stock price by studying stock price patterns using various machine learning techniques. This study suggest a way of predicting stock price patterns based on the Convolutional Neural Network(CNN) among machine learning techniques. CNN uses neural networks to classify images by extracting features from images through convolutional layers. Therefore, this study tries to classify candlestick images made by stock data in order to predict patterns. This study has two objectives. The first one referred as Case 1 is to predict the patterns with the images made by the same-day stock price data. The second one referred as Case 2 is to predict the next day stock price patterns with the images produced by the daily stock price data. In Case 1, data augmentation methods - random modification and Gaussian noise - are applied to generate more training data, and the generated images are put into the model to fit. Given that deep learning requires a large amount of data, this study suggests a method of data augmentation for candlestick images. Also, this study compares the accuracies of the images with Gaussian noise and different classification problems. All data in this study is collected through OpenAPI provided by DaiShin Securities. Case 1 has five different labels depending on patterns. The patterns are up with up closing, up with down closing, down with up closing, down with down closing, and staying. The images in Case 1 are created by removing the last candle(-1candle), the last two candles(-2candles), and the last three candles(-3candles) from 60 minutes, 30 minutes, 10 minutes, and 5 minutes candle charts. 60 minutes candle chart means one candle in the image has 60 minutes of information containing an open price, high price, low price, close price. Case 2 has two labels that are up and down. This study for Case 2 has generated for 60 minutes, 30 minutes, 10 minutes, and 5minutes candle charts without removing any candle. Considering the stock data, moving the candles in the images is suggested, instead of existing data augmentation techniques. How much the candles are moved is defined as the modified value. The average difference of closing prices between candles was 0.0029. Therefore, in this study, 0.003, 0.002, 0.001, 0.00025 are used for the modified value. The number of images was doubled after data augmentation. When it comes to Gaussian Noise, the mean value was 0, and the value of variance was 0.01. For both Case 1 and Case 2, the model is based on VGG-Net16 that has 16 layers. As a result, 10 minutes -1candle showed the best accuracy among 60 minutes, 30 minutes, 10 minutes, 5minutes candle charts. Thus, 10 minutes images were utilized for the rest of the experiment in Case 1. The three candles removed from the images were selected for data augmentation and application of Gaussian noise. 10 minutes -3candle resulted in 79.72% accuracy. The accuracy of the images with 0.00025 modified value and 100% changed candles was 79.92%. Applying Gaussian noise helped the accuracy to be 80.98%. According to the outcomes of Case 2, 60minutes candle charts could predict patterns of tomorrow by 82.60%. To sum up, this study is expected to contribute to further studies on the prediction of stock price patterns using images. This research provides a possible method for data augmentation of stock data.