• 한국멀티미디어학회논문지
• /
• 제21권10호
• /
• pp.1142-1149
• /
• 2018

Recently, computer vision application is increasing by using CNN which is one of the deep learning algorithms. However, CNN does not provide perfect classification performance due to gradient vanishing problem. Most of CNN algorithms use an activation function called ReLU to mitigate the gradient vanishing problem. In this study, four activation functions that can replace ReLU were applied to four different structural networks. Experimental results show that ReLU has the lowest performance in accuracy, loss rate, and speed of initial learning convergence from 20 experiments. It is concluded that the optimal activation function varied from network to network but the four activation functions were higher than ReLU.

• 전기학회논문지
• /
• 제67권11호
• /
• pp.1536-1541
• /
• 2018

For effective electricity consumption in urban railway station such as peak load shaving, it is important to know each electrical load pattern by various usage. The total electricity consumption in the urban railway substation is already measured in Korea, but the electricity consumption for each usage is not measured. The author proposed the deep learning method to estimate the electrical load pattern for each usage in the urban railway substation with public data such as weather data. GRU (gated recurrent unit), a variation on the LSTM (long short-term memory), was used, which aims to solve the vanishing gradient problem of standard a RNN (recursive neural networks). The optimal model was found and the estimation results with that were assessed.

• 한국어정보학회:학술대회논문집
• /
• 한국어정보학회 2017년도 제29회 한글및한국어정보처리학술대회
• /
• pp.238-241
• /
• 2017

본 논문에서는 구글 학술 검색 기반의 데이터를 이용하여 질병과 폐질환과 관련된 바이오마커 단어의 유사도를 계산하는 방법을 제안한다. 질병과 바이오마커의 유사도를 계산할 때, 각 단어의 구글 학술 검색의 검색 결과를 이용하였다. 이를 통해 폐질환 관련 바이오마커와 다른 질병간의 관계를 파악하고자 히며, 의료 전문가에게 폐질환 관련 바이오마커와 다른 질병간의 새로운 관계를 제시하고자 한다. 이러한 데이터를 이용하여 계산한 결과, Wor2Vec의 결과를 이용한 코사인 유사도의 결과와 상관 계수가 약 0.64로 상당히 높은 상관 관계를 확인할 수 있었다. 따라서 이 방법을 통해 질병과 바이오마커의 관계를 파악하고자 하였다. 또한 Word2Vec을 이용한 질병과 바이오마커 단어의 벡터 값과 단어 유사도 계산 방법의 결과를 이용한 Deep Neural Networks (DNNs) 모델을 구축하고자 하며, 이를 통해 자동적으로 유사도를 분석하고자 하였다.

• IEIE Transactions on Smart Processing and Computing
• /
• 제6권1호
• /
• pp.53-59
• /
• 2017

Optical character recognition (OCR) automatically recognizes text in an image. OCR is still a challenging problem in computer vision. A successful solution to OCR has important device applications, such as text-to-speech conversion and automatic document classification. In this work, we analyze character recognition performance using the current state-of-the-art deep-learning structures. One is the AlexNet structure, another is the LeNet structure, and the other one is the SPNet structure. For this, we have built our own dataset that contains digits and upper- and lower-case characters. We experiment in the presence of salt-and-pepper noise or Gaussian noise, and report the performance comparison in terms of recognition error. Experimental results indicate by five-fold cross-validation that the SPNet structure (our approach) outperforms AlexNet and LeNet in recognition error.

• International Journal of Contents
• /
• 제17권3호
• /
• pp.1-14
• /
• 2021

Since the outbreak of the coronavirus disease 2019 (COVID-19) pandemic, a lot of efforts have been made in the field of data science to help combat against this disease. Among them, forecasting the number of cases of infection is a crucial problem to predict the development of the pandemic. Many deep learning-based models can be applied to solve this type of time series problem. In this research, we would like to take a step forward to incorporate spatial data (geography) with time series data to forecast the cases of region-level infection simultaneously. Specifically, we model a single spatio-temporal graph, in which nodes represent the geographic regions, spatial edges represent the distance between each pair of regions, and temporal edges indicate the node features through time. We evaluate this approach in COVID-19 in a Korean dataset, and we show a decrease of approximately 10% in both RMSE and MAE, and a significant boost to the training speed compared to the baseline models. Moreover, the training efficiency allows this approach to be extended for a large-scale spatio-temporal dataset.

• ETRI Journal
• /
• 제41권1호
• /
• pp.109-116
• /
• 2019

This paper explores new techniques that are based on a hidden-layer linear transformation for fast speaker adaptation used in deep neural networks (DNNs). Conventional methods using affine transformations are ineffective because they require a relatively large number of parameters to perform. Meanwhile, methods that employ singular-value decomposition (SVD) are utilized because they are effective at reducing adaptive parameters. However, a matrix decomposition is computationally expensive when using online services. We propose the use of an extended diagonal linear transformation method to minimize adaptation parameters without SVD to increase the performance level for tasks that require smaller degrees of adaptation. In Korean large vocabulary continuous speech recognition (LVCSR) tasks, the proposed method shows significant improvements with error-reduction rates of 8.4% and 17.1% in five and 50 conversational sentence adaptations, respectively. Compared with the adaptation methods using SVD, there is an increased recognition performance with fewer parameters.

• Smart Structures and Systems
• /
• 제23권4호
• /
• pp.393-403
• /
• 2019

Applying more features gives us better accuracy in modeling; however, increasing the inputs causes the curse of dimensions. In this paper, a new structure has been proposed for fault detecting and identifying (FDI) of high-dimensional systems. This structure consist of two structure. The first part includes Auto-Encoders (AE) as Deep Neural Networks (DNNs) to produce feature engineering process and summarize the features. The second part consists of the Local Model Networks (LMNs) with LOcally LInear MOdel Tree (LOLIMOT) algorithm to model outputs (multiple models). The fault detection is based on these multiple models. Hence the residuals generated by comparing the system output and multiple models have been used to alarm the faults. To show the effectiveness of the proposed structure, it is tested on single-shaft industrial gas turbine prototype model. Finally, a brief comparison between the simulated results and several related works is presented and the well performance of the proposed structure has been illustrated.

• International Journal of Internet, Broadcasting and Communication
• /
• 제13권1호
• /
• pp.100-106
• /
• 2021

Over the past few decades, natural language processing research has not made much. However, the widespread use of deep learning and neural networks attracted attention for the application of neural networks in natural language processing. Sentiment analysis is one of the challenges of natural language processing. Emotions are things that a person thinks and feels. Therefore, sentiment analysis should be able to analyze the person's attitude, opinions, and inclinations in text or actual text. In the case of emotion analysis, it is a priority to simply classify two emotions: positive and negative. In this paper we propose the deep learning based sentimental analysis system according to various optimizer that is SGD, ADAM and RMSProp. Through experimental result RMSprop optimizer shows the best performance compared to others on IMDB data set. Future work is to find more best hyper parameter for sentimental analysis system.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권4호
• /
• pp.1464-1485
• /
• 2021

Emotional style of multimedia art works are abstract content information. This study aims to explore emotional style transfer method and find the possible way of matching music with appropriate images in respect to emotional style. DCNNs (Deep Convolutional Neural Networks) can capture style and provide emotional style transfer iterative solution for affective image generation. Here, we learn the image emotion features via DCNNs and map the affective style on the other images. We set image emotion feature as the style target in this style transfer problem, and held experiments to handle affective image generation of eight emotion categories, including dignified, dreaming, sad, vigorous, soothing, exciting, joyous, and graceful. A user study was conducted to test the synesthesia emotional image style transfer result with ground truth user perception triggered by the music-image pairs' stimuli. The transferred affective image result for music-image emotional synesthesia perception was proved effective according to user study result.

• 반도체디스플레이기술학회지
• /
• 제21권3호
• /
• pp.7-11
• /
• 2022

Object detection technology is one of the main research topics in the field of computer vision and has established itself as an essential base technology for implementing various vision systems. Recent DNN (Deep Neural Networks)-based algorithms achieve much higher recognition accuracy than traditional algorithms. However, it is well-known that the DNN model inference operation requires a relatively high computational power. In this paper, we analyze the inference time complexity of the state-of-the-art object detection architecture Yolov7 in various environments. Specifically, we compare and analyze the time complexity of four types of the Yolov7 model, YOLOv7-tiny, YOLOv7, YOLOv7-X, and YOLOv7-E6 when performing inference operations using CPU and GPU. Furthermore, we analyze the time complexity variation when inferring the same models using the Pytorch framework and the Onnxruntime engine.