• Journal of Intelligence and Information Systems
• /
• v.28 no.1
• /
• pp.89-106
• /
• 2022

Recently, the amount of video data collected from smartphones, CCTVs, black boxes, and high-definition cameras has increased rapidly. According to the increasing video data, the requirements for analysis and utilization are increasing. Due to the lack of skilled manpower to analyze videos in many industries, machine learning and artificial intelligence are actively used to assist manpower. In this situation, the demand for various computer vision technologies such as object detection and tracking, action detection, emotion detection, and Re-ID also increased rapidly. However, the object detection and tracking technology has many difficulties that degrade performance, such as re-appearance after the object's departure from the video recording location, and occlusion. Accordingly, action and emotion detection models based on object detection and tracking models also have difficulties in extracting data for each object. In addition, deep learning architectures consist of various models suffer from performance degradation due to bottlenects and lack of optimization. In this study, we propose an video analysis system consists of YOLOv5 based DeepSORT object tracking model, SlowFast based action recognition model, Torchreid based Re-ID model, and AWS Rekognition which is emotion recognition service. Proposed model uses single-linkage hierarchical clustering based Re-ID and some processing method which maximize hardware throughput. It has higher accuracy than the performance of the re-identification model using simple metrics, near real-time processing performance, and prevents tracking failure due to object departure and re-emergence, occlusion, etc. By continuously linking the action and facial emotion detection results of each object to the same object, it is possible to efficiently analyze videos. The re-identification model extracts a feature vector from the bounding box of object image detected by the object tracking model for each frame, and applies the single-linkage hierarchical clustering from the past frame using the extracted feature vectors to identify the same object that failed to track. Through the above process, it is possible to re-track the same object that has failed to tracking in the case of re-appearance or occlusion after leaving the video location. As a result, action and facial emotion detection results of the newly recognized object due to the tracking fails can be linked to those of the object that appeared in the past. On the other hand, as a way to improve processing performance, we introduce Bounding Box Queue by Object and Feature Queue method that can reduce RAM memory requirements while maximizing GPU memory throughput. Also we introduce the IoF(Intersection over Face) algorithm that allows facial emotion recognized through AWS Rekognition to be linked with object tracking information. The academic significance of this study is that the two-stage re-identification model can have real-time performance even in a high-cost environment that performs action and facial emotion detection according to processing techniques without reducing the accuracy by using simple metrics to achieve real-time performance. The practical implication of this study is that in various industrial fields that require action and facial emotion detection but have many difficulties due to the fails in object tracking can analyze videos effectively through proposed model. Proposed model which has high accuracy of retrace and processing performance can be used in various fields such as intelligent monitoring, observation services and behavioral or psychological analysis services where the integration of tracking information and extracted metadata creates greate industrial and business value. In the future, in order to measure the object tracking performance more precisely, there is a need to conduct an experiment using the MOT Challenge dataset, which is data used by many international conferences. We will investigate the problem that the IoF algorithm cannot solve to develop an additional complementary algorithm. In addition, we plan to conduct additional research to apply this model to various fields' dataset related to intelligent video analysis.

• Convergence Security Journal
• /
• v.21 no.2
• /
• pp.57-66
• /
• 2021

Deep neural networks (DNNs) provide excellent performance for image, speech, and pattern recognition. However, DNNs sometimes misrecognize certain adversarial examples. An adversarial example is a sample that adds optimized noise to the original data, which makes the DNN erroneously misclassified, although there is nothing wrong with the human eye. Therefore studies on defense against adversarial example attacks are required. In this paper, we have experimentally analyzed the success rate of detection for adversarial examples by adjusting various parameters. The performance of the ensemble defense method was analyzed using fast gradient sign method, DeepFool method, Carlini & Wanger method, which are adversarial example attack methods. Moreover, we used MNIST as experimental data and Tensorflow as a machine learning library. As an experimental method, we carried out performance analysis based on three adversarial example attack methods, threshold, number of models, and random noise. As a result, when there were 7 models and a threshold of 1, the detection rate for adversarial example is 98.3%, and the accuracy of 99.2% of the original sample is maintained.

• Knowledge Management Research
• /
• v.22 no.2
• /
• pp.1-17
• /
• 2021

In this study, we predicted the bitcoin prices of Bithum and Coinbase, a leading exchange in Korea and USA, using ARIMA and Recurrent Neural Networks(RNNs). And we used news articles from each country to suggest a separated RNN model. The suggested model identifies the datasets based on the changing trend of prices in the training data, and then applies time series prediction technique(RNNs) to create multiple models. Then we used daily news data to create a term-based dictionary for each trend change point. We explored trend change points in the test data using the daily news keyword data of testset and term-based dictionary, and apply a matching model to produce prediction results. With this approach we obtained higher accuracy than the model which predicted price by applying just time series prediction technique. This study presents that the limitations of the time series prediction techniques could be overcome by exploring trend change points using news data and various time series prediction techniques with text mining techniques could be applied to improve the performance of the model in the further research.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.8
• /
• pp.572-578
• /
• 2020

The stock market is affected by unexpected factors, such as politics, society, and natural disasters, as well as by corporate performance and economic conditions. In recent days, artificial intelligence has become popular, and many researchers have tried to conduct experiments with that. Our study proposes an experiment using not only stock-related data but also other various economic data. We acquired a year's worth of data on stock prices, the percentage of foreigners, interest rates, and exchange rates, and combined them in various ways. Thus, our input data became diversified, and we put the combined input data into a nonlinear autoregressive network with exogenous inputs (NARX) model. With the input data in the NARX model, we analyze and compare them to the original data. As a result, the model exhibits a root mean square error (RMSE) of 0.08 as being the most accurate when we set 10 neurons and two delays with a combination of stock prices and exchange rates from the U.S., China, Europe, and Japan. This study is meaningful in that the exchange rate has the greatest influence on stock prices, lowering the error from RMSE 0.589 when only closing data are used.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.10
• /
• pp.387-396
• /
• 2018

Stock price prediction has been a difficult problem to solve. There have been many studies to predict stock price scientifically, but it is still impossible to predict the exact price. Recently, a variety of types of cryptocurrency has been developed, beginning with Bitcoin, which is technically implemented as the concept of distributed ledger. Various approaches have been attempted to predict the price of cryptocurrency. Especially, it is various from attempts to stock prediction techniques in traditional stock market, to attempts to apply deep learning and reinforcement learning. Since the market for cryptocurrency has many new features that are not present in the existing traditional stock market, there is a growing demand for new analytical techniques suitable for the cryptocurrency market. In this study, we first collect and process seven cryptocurrency price data through Bithumb's API. Then, we use the gradient boosting model, which is a data-driven learning based machine learning model, and let the model learn the price data change of cryptocurrency. We also find the most optimal model parameters in the verification step, and finally evaluate the prediction performance of the cryptocurrency price trends.

• The Journal of Society for e-Business Studies
• /
• v.22 no.4
• /
• pp.71-101
• /
• 2017

Since late 2016, protectionism has been a major trend in world trade with the Great Britain exiting the European Union and the United States electing Donald Trump as the 45th president. Consequently, there has been a huge public outcry regarding the negative prospects of heavy industry firms in Korea, which are highly dependent upon international trade with Western countries including the United States. In light of such trend and concerns, we have tried to predict business performance of heavy industry firms in Korea with data regarding trade policy of the United States. United States International Trade Commission (USITC) levies countervailing duties and anti-dumping duties to firms that violate its fair-trade regulations. In this study, we have performed data analysis with past records of countervailing duties and anti-dumping duties. With results from clustering analysis, it could be concluded that trade policy trends of the Unites States significantly affects the business performance of heavy industry firms in Korea. Furthermore, we have attempted to quantify such effects by employing long short-term memory (LSTM), a popular neural networks model that is well-suited to deal with sequential data. Our major contribution is that we have succeeded in empirically validating the intuitive argument and also predicting the future trend with rigorous data mining techniques. With some improvements, our results are expected to be highly relevant to designing regulations regarding heavy industry in Korea.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.22 no.3
• /
• pp.239-248
• /
• 2020

Exact rock classification helps suitable support patterns to be installed. Face mapping is usually conducted to classify the rock mass using RMR (Rock Mass Ration) or Q values. There have been several attempts to predict the grade of rock mass using mechanical data of jumbo drills or probe drills and photographs of excavation surfaces by using deep learning. However, they took long time, or had a limitation that it is impossible to grasp the rock grade in ahead of the tunnel surface. In this study, a method to predict the Q value ahead of excavation surface is developed using recurrent neural network (RNN) technique and it is compared with the Q values from face mapping for verification. Among Q values from over 4,600 tunnel faces, 70% of data was used for learning, and the rests were used for verification. Repeated learnings were performed in different number of learning and number of previous excavation surfaces utilized for learning. The coincidence between the predicted and actual Q values was compared with the root mean square error (RMSE). RMSE value from 600 times repeated learning with 2 prior excavation faces gives a lowest values. The results from this study can vary with the input data sets, the results can help to understand how the past ground conditions affect the future ground conditions and to predict the Q value ahead of the tunnel excavation face.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.4
• /
• pp.165-172
• /
• 2020

In this paper we propose an uses on-device-based edge computing technology and big data analysis methods through the use of on-device-based edge computing technology and analysis of big data, which are distributed computing paradigms that introduce computations and storage devices where necessary to solve problems such as transmission delays that occur when data is transmitted to central centers and processed in current general smart factories. However, even if edge computing-based technology is applied in practice, the increase in devices on the network edge will result in large amounts of data being transferred to the data center, resulting in the network band reaching its limits, which, despite the improvement of network technology, does not guarantee acceptable transfer speeds and response times, which are critical requirements for many applications. It provides the basis for developing into an AI-based facility prediction conservation analysis tool that can apply deep learning suitable for big data in the future by supporting intelligent facility management that can support productivity growth through research that can be applied to the field of facility preservation and smart factory industry with integrated hardware technology that can accommodate these requirements and factory management and control technology.

• Journal of Convergence for Information Technology
• /
• v.10 no.12
• /
• pp.201-207
• /
• 2020

In this study, the practicality of unmanned aerial vehicle photography information was identified. Therefore, a total of four consecutive surveys were conducted on the field-level survey areas among the areas subject to photography using unmanned aerial vehicles, and the changes in crop conditions were analyzed using pictures of unmanned aerial vehicles taken during each survey. It is appropriate to collect and utilize photographic information by directly taking pictures of the survey area according to the time of the on-site survey using unmanned aerial vehicles in the field layer, which is an area where many changes in topography, crop vegetation, and crop types are expected. And it turned out that it was appropriate to utilize satellite images in consideration of economic and efficient aspects in relatively unchanged rice paddies and facilities. If the survey area is well equipped with systems for crop cultivation, deep learning can be utilized in real time by utilizing libraries after obtaining photographic data for a certain area using unmanned aircraft in the future. Through this process, it is believed that it can be used to analyze the overall crop and shipment volume by identifying the crop status and surveying the quantity per unit area.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.399-401
• /
• 2022

A definitive diagnosis to classify the breast malignancy status may be achieved by microscopic analysis using surgical open biopsy. However, this procedure requires experts in the specializing of histopathological image analysis directing to time-consuming and high cost. To overcome these issues, deep learning is considered practically efficient to categorize breast cancer into benign and malignant from histopathological images in order to assist pathologists. This study presents a pre-trained convolutional neural network model architecture with a 100% fine-tuning scheme and Adagrad optimizer to classify the breast cancer histopathological images into benign and malignant using a 40× magnification BreaKHis dataset. The pre-trained architecture was constructed using the InceptionResNetV2 model to generate a modified InceptionResNetV2 by substituting the last layer with dense and dropout layers. The results by demonstrating training loss of 0.25%, training accuracy of 99.96%, validation loss of 3.10%, validation accuracy of 99.41%, test loss of 8.46%, and test accuracy of 98.75% indicated that the modified InceptionResNetV2 model is reliable to predict the breast malignancy type from histopathological images. Future works are necessary to focus on k-fold cross-validation, optimizer, model, hyperparameter optimization, and classification on 100×, 200×, and 400× magnification.