• Journal of Broadcast Engineering
• /
• v.24 no.2
• /
• pp.234-242
• /
• 2019

In this paper, we propose a deep neural network which extracts SIFT feature points by determining whether the center pixel of a cropped image is a SIFT feature point. The data set of this network consists of a DIV2K dataset cut into $33{\times}33$ size and uses RGB image unlike SIFT which uses black and white image. The ground truth consists of the RobHess SIFT features extracted by setting the octave (scale) to 0, the sigma to 1.6, and the intervals to 3. Based on the VGG-16, we construct an increasingly deep network of 13 to 23 and 33 convolution layers, and experiment with changing the method of increasing the image scale. The result of using the sigmoid function as the activation function of the output layer is compared with the result using the softmax function. Experimental results show that the proposed network not only has more than 99% extraction accuracy but also has high extraction repeatability for distorted images.

• 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 the Korea Society of Computer and Information
• /
• v.25 no.9
• /
• pp.31-36
• /
• 2020

This research study designed a location image collecting technology. It provides the exact location information of an image which is not given in the photo to the user. Deep learning technology analysis and collects the images. The purpose of this service system is to provide the exact place name, location and the various information of the place such as nearby recommended attractions when the user upload the image photo to the service system. Suggested system has a deep learning model that has a size of 25.3MB, and the model repeats the learning process 50 times with a total of 15,266 data, performing 93.75% of the final accuracy. This system can also be linked with various services potentially for further development.

• Journal of Digital Convergence
• /
• v.19 no.12
• /
• pp.353-358
• /
• 2021

Parkinson's disease is the second most common degenerative brain disease after Alzheimer's in old age. Symptoms of Parkinson's disease are factors that reduce the quality of life in daily life, such as shaking hands, slowing behavior and cognitive function. Parkinson's disease that can slow the progression of the disease through early diagnosis. To diagnoze Parkinson's disease early, an algorithm was implemented to extract features using wav2vec and to diagnose the presence or absence of Parkinson's disease with deep learning(ANN). As a results of the experiment, the accuracy was 97.47%. It was better than the results of diagnosing Parkinson's disease using the existing neural network. The audio voice file could simply reduce the experiment process and obtain improved results.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.15 no.6
• /
• pp.7-14
• /
• 2019

The automatic monitoring and detection of crowd behavior in the surveillance videos has obtained significant attention in the field of computer vision due to its vast applications such as security, safety and protection of assets etc. Also, the field of crowd analysis is growing upwards in the research community. For this purpose, it is very necessary to detect and analyze the crowd behavior. In this paper, we proposed a deep learning-based method which detects abnormal activities in surveillance cameras installed in a smart city. A fine-tuned VGG-16 model is trained on publicly available benchmark crowd dataset and is tested on real-time streaming. The CCTV camera captures the video stream, when abnormal activity is detected, an alert is generated and is sent to the nearest police station to take immediate action before further loss. We experimentally have proven that the proposed method outperforms over the existing state-of-the-art techniques.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.5
• /
• pp.53-59
• /
• 2018

Due to the recent developments in computer vision technology, complex actions can be recognized with reasonable accuracy in smart cities. In contrast, violence recognition such as events related to fight and knife, has gained less attention. The capability of visual surveillance can be used for detecting fight in streets or in prison centers. In this paper, we proposed a deep learning-based violence recognition method for surveillance cameras. A convolutional neural network (CNN) model is trained and fine-tuned on available benchmark datasets of fights and knives for violence recognition. When an abnormal event is detected, an alarm can be sent to the nearest police station to take immediate action. Moreover, when the probabilities of fight and knife classes are predicted very low, this situation is considered as normal situation. The experimental results of the proposed method outperformed other state-of-the-art CNN models with high margin by achieving maximum 99.21% accuracy.

• Electronics and Telecommunications Trends
• /
• v.33 no.4
• /
• pp.14-22
• /
• 2018

Deep neural networks (DNNs) are widely used in various domains such as speech and image recognition. DNN software frameworks such as Tensorflow and Caffe contributed to the popularity of DNN because of their easy programming environment. In addition, many companies are developing neuromorphic processing units (NPU) such as Tensor Processing Units (TPUs) and Graphical Processing Units (GPUs) to improve the performance of DNN processing. However, there is a large gap between NPUs and DNN software frameworks due to the lack of framework support for various NPUs. A bridge for the gap is a DNN software platform including DNN optimized compilers and DNN libraries. In this paper, we review the technical trends of DNN software platforms.

• Journal of Korea Multimedia Society
• /
• v.22 no.7
• /
• pp.749-758
• /
• 2019

This paper proposes a data augmentation algorithm to improve the performance of DNN(Deep Neural Network) based speech enhancement. Many deep learning models are exploring algorithms to maximize the performance in limited amount of data. The most commonly used algorithm is the data augmentation which is the technique artificially increases the amount of data. For the effective data augmentation algorithm, we used a formant enhancement method that assign the different weights to the formant frequencies. The DNN model which is trained using the proposed data augmentation algorithm was evaluated in various noise environments. The speech enhancement performance of the DNN model with the proposed data augmentation algorithm was compared with the algorithms which are the DNN model with the conventional data augmentation and without the data augmentation. As a result, the proposed data augmentation algorithm showed the higher speech enhancement performance than the other algorithms.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.327-328
• /
• 2020

본 논문에서는 저전력 저비용 임베디드 환경에서 PCB 검사 기법을 제안한다. 특히, IC 미삽에 대한 검출 알고리즘을 제안하고 실험한다. 고사양의 컴퓨팅 시스템에서는 CNN과 같은 딥러닝 뉴럴 네트워크를 사용하여 특별한 알고리즘을 고려하지 않아도 대규모의 데이터를 입력함으로써 모델을 완성하고 이를 통해 PCB 검사를 수행할 수 있다. 그러나 데이터의 양이 충분하지 않거나 충분한 전력과 비용을 투입하지 못하는 임베디드 환경에서는 각 부품에 따른 컴퓨터 비전 알고리즘이 필요하다. IC의 경우 타부품에 비하여 형태가 직사각으로 정형화 되있으며 색상도 균일한 특징을 가지고 있기에 미삽에 대한 검출이 가능하다. 베어보드(Bare Board)의 색상과 IC 부품의 색상이 확연히 다를 경우에는 RGB 픽셀을 카운트 하는 히스토그램 카운팅 알고리즘만으로 검출이 가능하다. 베어보드의 색삭과 IC의 색상이 유사할 경우에는 베어보드의 핀 혹은 홀의 형태를 감지하여 검출이 가능하다. 본 논문에서는 베어보드의 색상와 IC의 색상이 같을 경우에 다를 경우를 나누어 미삽 검사를 수행하고 그 정확도를 확인한다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.675-679
• /
• 2020

최근 이미지 분류의 성능 향상을 위해 깊은 레이어와 넓은 채널을 가지는 모델들이 제안되어져 왔다. 높은 분류 정확도를 보이는 모델을 제안하는 것은 과한 컴퓨팅 파워와 계산시간을 요구한다. 본 논문에서는 이미지 분류 기법에서 사용되는 딥 뉴럴 네트워크 모델에 있어, 프루닝 방법을 통해 상대적으로 불필요한 가중치를 제거함과 동시에 분류 정확도 하락을 최소로 하는 동적 필터 프루닝 방법을 제시한다. 원샷 프루닝 기법, 정적 필터 프루닝 기법과 다르게 제거된 가중치에 대해서 소생 기회를 제공함으로써 더 좋은 성능을 보인다. 또한, 재학습이 필요하지 않기 때문에 빠른 계산 속도와 적은 컴퓨팅 파워를 보장한다. ResNet20 에서 CIFAR10 데이터셋에 대하여 실험한 결과 약 50%의 압축률에도 88.74%의 분류 정확도를 보였다.