• Journal of Internet Computing and Services
• /
• v.21 no.6
• /
• pp.23-31
• /
• 2020

This study was carried out to generate various images of railroad surfaces with random defects as training data to be better at the detection of defects. Defects on the surface of railroads are caused by various factors such as friction between track binding devices and adjacent tracks and can cause accidents such as broken rails, so railroad maintenance for defects is necessary. Therefore, various researches on defect detection and inspection using image processing or machine learning on railway surface images have been conducted to automate railroad inspection and to reduce railroad maintenance costs. In general, the performance of the image processing analysis method and machine learning technology is affected by the quantity and quality of data. For this reason, some researches require specific devices or vehicles to acquire images of the track surface at regular intervals to obtain a database of various railway surface images. On the contrary, in this study, in order to reduce and improve the operating cost of image acquisition, we constructed the 'Defective Railroad Surface Regeneration Model' by applying the methods presented in the related studies of the Generative Adversarial Network (GAN). Thus, we aimed to detect defects on railroad surface even without a dedicated database. This constructed model is designed to learn to generate the railroad surface combining the different railroad surface textures and the original surface, considering the ground truth of the railroad defects. The generated images of the railroad surface were used as training data in defect detection network, which is based on Fully Convolutional Network (FCN). To validate its performance, we clustered and divided the railroad data into three subsets, one subset as original railroad texture images and the remaining two subsets as another railroad surface texture images. In the first experiment, we used only original texture images for training sets in the defect detection model. And in the second experiment, we trained the generated images that were generated by combining the original images with a few railroad textures of the other images. Each defect detection model was evaluated in terms of 'intersection of union(IoU)' and F1-score measures with ground truths. As a result, the scores increased by about 10~15% when the generated images were used, compared to the case that only the original images were used. This proves that it is possible to detect defects by using the existing data and a few different texture images, even for the railroad surface images in which dedicated training database is not constructed.

• Journal of Intelligence and Information Systems
• /
• v.25 no.4
• /
• pp.141-154
• /
• 2019

Rapid growth of internet technology and social media is progressing. Data mining technology has evolved to enable unstructured document representations in a variety of applications. Sentiment analysis is an important technology that can distinguish poor or high-quality content through text data of products, and it has proliferated during text mining. Sentiment analysis mainly analyzes people's opinions in text data by assigning predefined data categories as positive and negative. This has been studied in various directions in terms of accuracy from simple rule-based to dictionary-based approaches using predefined labels. In fact, sentiment analysis is one of the most active researches in natural language processing and is widely studied in text mining. When real online reviews aren't available for others, it's not only easy to openly collect information, but it also affects your business. In marketing, real-world information from customers is gathered on websites, not surveys. Depending on whether the website's posts are positive or negative, the customer response is reflected in the sales and tries to identify the information. However, many reviews on a website are not always good, and difficult to identify. The earlier studies in this research area used the reviews data of the Amazon.com shopping mal, but the research data used in the recent studies uses the data for stock market trends, blogs, news articles, weather forecasts, IMDB, and facebook etc. However, the lack of accuracy is recognized because sentiment calculations are changed according to the subject, paragraph, sentiment lexicon direction, and sentence strength. This study aims to classify the polarity analysis of sentiment analysis into positive and negative categories and increase the prediction accuracy of the polarity analysis using the pretrained IMDB review data set. First, the text classification algorithm related to sentiment analysis adopts the popular machine learning algorithms such as NB (naive bayes), SVM (support vector machines), XGboost, RF (random forests), and Gradient Boost as comparative models. Second, deep learning has demonstrated discriminative features that can extract complex features of data. Representative algorithms are CNN (convolution neural networks), RNN (recurrent neural networks), LSTM (long-short term memory). CNN can be used similarly to BoW when processing a sentence in vector format, but does not consider sequential data attributes. RNN can handle well in order because it takes into account the time information of the data, but there is a long-term dependency on memory. To solve the problem of long-term dependence, LSTM is used. For the comparison, CNN and LSTM were chosen as simple deep learning models. In addition to classical machine learning algorithms, CNN, LSTM, and the integrated models were analyzed. Although there are many parameters for the algorithms, we examined the relationship between numerical value and precision to find the optimal combination. And, we tried to figure out how the models work well for sentiment analysis and how these models work. This study proposes integrated CNN and LSTM algorithms to extract the positive and negative features of text analysis. The reasons for mixing these two algorithms are as follows. CNN can extract features for the classification automatically by applying convolution layer and massively parallel processing. LSTM is not capable of highly parallel processing. Like faucets, the LSTM has input, output, and forget gates that can be moved and controlled at a desired time. These gates have the advantage of placing memory blocks on hidden nodes. The memory block of the LSTM may not store all the data, but it can solve the CNN's long-term dependency problem. Furthermore, when LSTM is used in CNN's pooling layer, it has an end-to-end structure, so that spatial and temporal features can be designed simultaneously. In combination with CNN-LSTM, 90.33% accuracy was measured. This is slower than CNN, but faster than LSTM. The presented model was more accurate than other models. In addition, each word embedding layer can be improved when training the kernel step by step. CNN-LSTM can improve the weakness of each model, and there is an advantage of improving the learning by layer using the end-to-end structure of LSTM. Based on these reasons, this study tries to enhance the classification accuracy of movie reviews using the integrated CNN-LSTM model.

• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.1-25
• /
• 2020

In this paper, we suggest an application system architecture which provides accurate, fast and efficient automatic gasometer reading function. The system captures gasometer image using mobile device camera, transmits the image to a cloud server on top of private LTE network, and analyzes the image to extract character information of device ID and gas usage amount by selective optical character recognition based on deep learning technology. In general, there are many types of character in an image and optical character recognition technology extracts all character information in an image. But some applications need to ignore non-of-interest types of character and only have to focus on some specific types of characters. For an example of the application, automatic gasometer reading system only need to extract device ID and gas usage amount character information from gasometer images to send bill to users. Non-of-interest character strings, such as device type, manufacturer, manufacturing date, specification and etc., are not valuable information to the application. Thus, the application have to analyze point of interest region and specific types of characters to extract valuable information only. We adopted CNN (Convolutional Neural Network) based object detection and CRNN (Convolutional Recurrent Neural Network) technology for selective optical character recognition which only analyze point of interest region for selective character information extraction. We build up 3 neural networks for the application system. The first is a convolutional neural network which detects point of interest region of gas usage amount and device ID information character strings, the second is another convolutional neural network which transforms spatial information of point of interest region to spatial sequential feature vectors, and the third is bi-directional long short term memory network which converts spatial sequential information to character strings using time-series analysis mapping from feature vectors to character strings. In this research, point of interest character strings are device ID and gas usage amount. Device ID consists of 12 arabic character strings and gas usage amount consists of 4 ~ 5 arabic character strings. All system components are implemented in Amazon Web Service Cloud with Intel Zeon E5-2686 v4 CPU and NVidia TESLA V100 GPU. The system architecture adopts master-lave processing structure for efficient and fast parallel processing coping with about 700,000 requests per day. Mobile device captures gasometer image and transmits to master process in AWS cloud. Master process runs on Intel Zeon CPU and pushes reading request from mobile device to an input queue with FIFO (First In First Out) structure. Slave process consists of 3 types of deep neural networks which conduct character recognition process and runs on NVidia GPU module. Slave process is always polling the input queue to get recognition request. If there are some requests from master process in the input queue, slave process converts the image in the input queue to device ID character string, gas usage amount character string and position information of the strings, returns the information to output queue, and switch to idle mode to poll the input queue. Master process gets final information form the output queue and delivers the information to the mobile device. We used total 27,120 gasometer images for training, validation and testing of 3 types of deep neural network. 22,985 images were used for training and validation, 4,135 images were used for testing. We randomly splitted 22,985 images with 8:2 ratio for training and validation respectively for each training epoch. 4,135 test image were categorized into 5 types (Normal, noise, reflex, scale and slant). Normal data is clean image data, noise means image with noise signal, relfex means image with light reflection in gasometer region, scale means images with small object size due to long-distance capturing and slant means images which is not horizontally flat. Final character string recognition accuracies for device ID and gas usage amount of normal data are 0.960 and 0.864 respectively.

• 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.

• Journal of Intelligence and Information Systems
• /
• v.23 no.4
• /
• pp.147-168
• /
• 2017

There have been many studies on accurate stock market forecasting in academia for a long time, and now there are also various forecasting models using various techniques. Recently, many attempts have been made to predict the stock index using various machine learning methods including Deep Learning. Although the fundamental analysis and the technical analysis method are used for the analysis of the traditional stock investment transaction, the technical analysis method is more useful for the application of the short-term transaction prediction or statistical and mathematical techniques. Most of the studies that have been conducted using these technical indicators have studied the model of predicting stock prices by binary classification - rising or falling - of stock market fluctuations in the future market (usually next trading day). However, it is also true that this binary classification has many unfavorable aspects in predicting trends, identifying trading signals, or signaling portfolio rebalancing. In this study, we try to predict the stock index by expanding the stock index trend (upward trend, boxed, downward trend) to the multiple classification system in the existing binary index method. In order to solve this multi-classification problem, a technique such as Multinomial Logistic Regression Analysis (MLOGIT), Multiple Discriminant Analysis (MDA) or Artificial Neural Networks (ANN) we propose an optimization model using Genetic Algorithm as a wrapper for improving the performance of this model using Multi-classification Support Vector Machines (MSVM), which has proved to be superior in prediction performance. In particular, the proposed model named GA-MSVM is designed to maximize model performance by optimizing not only the kernel function parameters of MSVM, but also the optimal selection of input variables (feature selection) as well as instance selection. In order to verify the performance of the proposed model, we applied the proposed method to the real data. The results show that the proposed method is more effective than the conventional multivariate SVM, which has been known to show the best prediction performance up to now, as well as existing artificial intelligence / data mining techniques such as MDA, MLOGIT, CBR, and it is confirmed that the prediction performance is better than this. Especially, it has been confirmed that the 'instance selection' plays a very important role in predicting the stock index trend, and it is confirmed that the improvement effect of the model is more important than other factors. To verify the usefulness of GA-MSVM, we applied it to Korea's real KOSPI200 stock index trend forecast. Our research is primarily aimed at predicting trend segments to capture signal acquisition or short-term trend transition points. The experimental data set includes technical indicators such as the price and volatility index (2004 ~ 2017) and macroeconomic data (interest rate, exchange rate, S&P 500, etc.) of KOSPI200 stock index in Korea. Using a variety of statistical methods including one-way ANOVA and stepwise MDA, 15 indicators were selected as candidate independent variables. The dependent variable, trend classification, was classified into three states: 1 (upward trend), 0 (boxed), and -1 (downward trend). 70% of the total data for each class was used for training and the remaining 30% was used for verifying. To verify the performance of the proposed model, several comparative model experiments such as MDA, MLOGIT, CBR, ANN and MSVM were conducted. MSVM has adopted the One-Against-One (OAO) approach, which is known as the most accurate approach among the various MSVM approaches. Although there are some limitations, the final experimental results demonstrate that the proposed model, GA-MSVM, performs at a significantly higher level than all comparative models.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2017.04a
• /
• pp.142-142
• /
• 2017

본 연구에서는 파프리카를 자동 수확하기 위한 시스템 개발의 일환으로 파프리카 재배환경에서 획득한 영상 내에 존재하는 파프리카 영역과 비 파프리카 영역의 RGB 정보를 입력으로 하는 인공신경망을 설계하고 학습을 수행하고자 하였다. 학습된 신경망을 이용하여 영상 내 파프리카 영역과 비 파프리카 영역의 구분이 가능 할 것으로 사료된다. 심층 신경망을 설계하기 위하여 MS Visual studio 2015의 C++, MFC와 Python 및 TensorFlow를 사용하였다. 먼저, 심층 신경망은 입력층과 출력층, 그리고 은닉층 8개를 가지는 형태로 입력 뉴런 3개, 출력 뉴런 4개, 각 은닉층의 뉴런은 5개로 설계하였다. 일반적으로 심층 신경망에서는 은닉층이 깊을수록 적은 입력으로 좋은 학습 결과를 기대 할 수 있지만 소요되는 시간이 길고 오버 피팅이 일어날 가능성이 높아진다. 따라서 본 연구에서는 소요시간을 줄이기 위하여 Xavier 초기화를 사용하였으며, 오버 피팅을 줄이기 위하여 ReLU 함수를 활성화 함수로 사용하였다. 파프리카 재배환경에서 획득한 영상에서 파프리카 영역과 비 파프리카 영역의 RGB 정보를 추출하여 학습의 입력으로 하고 기대 출력으로 붉은색 파프리카의 경우 [0 0 1], 노란색 파프리카의 경우 [0 1 0], 비 파프리카 영역의 경우 [1 0 0]으로 하는 형태로 3538개의 학습 셋을 만들었다. 학습 후 학습 결과를 평가하기 위하여 30개의 테스트 셋을 사용하였다. 학습 셋을 이용하여 학습을 수행하기 위해 학습률을 변경하면서 학습 결과를 확인하였다. 학습률을 0.01 이상으로 설정한 경우 학습이 이루어지지 않았다. 이는 학습률에 의해 결정되는 가중치의 변화량이 너무 커서 비용 함수의 결과가 0에 수렴하지 않고 발산하는 경향에 의한 것으로 사료된다. 학습률을 0.005, 0.001로 설정 한 경우 학습에 성공하였다. 학습률 0.005의 경우 학습 횟수 3146회, 소요시간 20.48초, 학습 정확도 99.77%, 테스트 정확도 100%였으며, 학습률 0.001의 경우 학습 횟수 38931회, 소요시간 181.39초, 학습 정확도 99.95%, 테스트 정확도 100%였다. 학습률이 작을수록 더욱 정확한 학습이 가능하지만 소요되는 시간이 크고 국부 최소점에 빠질 확률이 높았다. 학습률이 큰 경우 학습 소요 시간이 줄어드는 반면 학습 과정에서 비용이 발산하여 학습이 이루어지지 않는 경우가 많음을 확인 하였다.

• Tunnel and Underground Space
• /
• v.17 no.4
• /
• pp.301-310
• /
• 2007

The prediction of temperature in the workings for the propriety examination for the development of a deep coal bed and the ventilation design is fairly important. It is quite demanding to obtain precise thermal conductivity of rock due to the variety and the complexity of the rock types contiguous to the coal bed. Therefore, to estimate the thermal conductivity corresponding to this geological situation and complex gallery conditions, a computing program which is TemPredict, is developed in this study. It employs Artificial Neural Network and calculates the climatic conditions in galleries. This advanced neural network is based upon the Back-Propagation Algorithm and composed of the input layers that are acceptant of the physical and geological factors of the coal bed and the hidden layers each of which has the 5 and 3 neurons. To verify TemPredict, the calculated result is compared with the measured one at the entrance of -300 ML 9X of Jang-sung production department, Jang-sung Coal Mine. The difference between the results calculated by TemPredict ($25.65^{\circ}C$) and measured ($25.7^{\circ}C$) is only $0.05^{\circ}C$, which is less than the allowable error 5%. The result has more than 95% of very high reliability. The temperature prediction for the main carriage gallery 9X in -425 ML under construction when it is completed is made. Its result is $28.2^{\circ}C$. In the future, it would contribute to the ventilation design for the mine and the underground structures.

• Journal of Broadcast Engineering
• /
• v.23 no.3
• /
• pp.421-430
• /
• 2018

Many voice recognition systems use methods such as MFCC, HMM to acknowledge human voice. This recognition method is designed to analyze only a targeted sound which normally appears between a human and a device one. However, the recognition capability is limited when there is a group sound formed with diversity in wider frequency range such as dog barking and indoor sounds. The frequency of overlapped sound resides in a wide range, up to 20KHz, which is higher than a voice. This paper proposes the new recognition method which provides wider frequency range by conjugating the Wideband Sound Spectrogram and the Keras Sequential Model based on DNN. The wideband sound spectrogram is adopted to analyze and verify diverse sounds from wide frequency range as it is designed to extract features and also classify as explained. The KSM is employed for the pattern recognition using extracted features from the WSS to improve sound recognition quality. The experiment verified that the proposed WSS and KSM excellently classified the targeted sound among noisy environment; overlapped sounds such as dog barking and indoor sounds. Furthermore, the paper shows a stage by stage analyzation and comparison of the factors' influences on the recognition and its characteristics according to various levels of noise.

• Cartoon and Animation Studies
• /
• s.40
• /
• pp.423-442
• /
• 2015

In a digital environment that is keep in change content from smart phones to tablet PC are a social network service is holding deep place in our lives. Social networking applications has building a network and communication with others than other application, it means that Social networking applications are sharing not only personal purpose in that trend of variety and competition of these social networks can be expected to trend and be developed thru analysis of user certification. this study of social network service application is proposed to developing of application thru analyze the two-effective application which is high ranked in google store. the theoretical foundation was set based on the seven elements of the social network service of the information structures designed by Jean Smith. This study proceeded analysis is for the functional elements of Facebook and Instagram, and the advantages and disadvantages through survey research. As a result of the empirical analysis to user of Instagram and face book of communication, identity, satisfaction for the group are equally. Instagram is about the presence, reputation, and Facebook has had a high level of satisfaction for each sharing and relationship. Facebook got high satisfaction from sharing features, but user feel of discomfort in the randomly showing advertising content. Instagram is not showing off advertise on common page of content, it is good point to be complementary to facebook. And, Instagram hashtag is good for convenience, but satisfaction is high with Facebook. in order to increase the satisfaction of Instagram, it is necessary to consider the main advantage of the communication and the functional aspects of the share from facebook.

• Journal of Biomedical Engineering Research
• /
• v.38 no.6
• /
• pp.321-329
• /
• 2017

This paper presents a bone metastasis Detection algorithm on abdominal computed tomography images for early detection using fully convolutional neural networks. The images were taken from patients with various cancers (such as lung cancer, breast cancer, colorectal cancer, etc), and thus the locations of those lesions were varied. To overcome the lack of data, we augmented the data by adjusting the brightness of the images or flipping the images. Before the augmentation, when 70% of the whole data were used in the pre-test, we could obtain the pixel-wise sensitivity of 18.75%, the specificity of 99.97% on the average of test dataset. With the augmentation, we could obtain the sensitivity of 30.65%, the specificity of 99.96%. The increase in sensitivity shows that the augmentation was effective. In the result obtained by using the whole data, the sensitivity of 38.62%, the specificity of 99.94% and the accuracy of 99.81% in the pixel-wise. lesion-wise sensitivity is 88.89% while the false alarm per case is 0.5. The results of this study did not reach the level that could substitute for the clinician. However, it may be helpful for radiologists when it can be used as a screening tool.