• The Journal of the Acoustical Society of Korea
• /
• v.39 no.4
• /
• pp.292-302
• /
• 2020

Machine learning algorithms have made immense contributions in various fields including sonar and radar applications. Recently developed Cycle-Consistency Generative Adversarial Network (CycleGAN), a variant of GAN has been successfully used for unpaired image-to-image translation. We present a modified CycleGAN for translation of underwater ship engine sounds with high perceptual quality. The proposed network is composed of an improved generator model trained to translate underwater audio from one vessel type to other, an improved discriminator to identify the data as real or fake and a modified cycle-consistency loss function. The quantitative and qualitative analysis of the proposed CycleGAN are performed on publicly available underwater dataset ShipsEar by evaluating and comparing Mel-cepstral distortion, pitch contour matching, nearest neighbor comparison and mean opinion score with existing algorithms. The analysis results of the proposed network demonstrate the effectiveness of the proposed network.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.9
• /
• pp.63-69
• /
• 2020

Image-image translation is a technology that creates a target image through input images, and has recently shown high performance in creating a more realistic image by utilizing GAN, which is a non-map learning structure. Therefore, there are various studies on image-to-image translation using GAN. At this point, most image-to-image translations basically target one attribute translation. But the data used and obtainable in real life consist of a variety of features that are hard to explain with one feature. Therefore, if you aim to change multiple attributes that can divide the image creation process by attributes to take advantage of the various attributes, you will be able to play a better role in image-to-image translation. In this paper, we propose Multi CycleGAN, a dual attribute transformation structure, by utilizing CycleGAN, which showed high performance among image-image translation structures using GAN. This structure implements a dual transformation structure in which three domains conduct two-way learning to learn about the two properties of an input domain. Experiments have shown that images through the new structure maintain the properties of the input area and show high performance with the target properties applied. Using this structure, it is possible to create more diverse images in the future, so we can expect to utilize image generation in more diverse areas.

• The Journal of Korea Robotics Society
• /
• v.13 no.1
• /
• pp.26-30
• /
• 2018

Planetary global localization is necessary for long-range rover missions in which communication with command center operator is throttled due to the long distance. There has been number of researches that address this problem by exploiting and matching rover surroundings with global digital elevation maps (DEM). Using conventional methods for matching, however, is challenging due to artifacts in both DEM rendered images, and/or rover 2D images caused by DEM low resolution, rover image illumination variations and small terrain features. In this work, we use train CNN discriminator to match rover 2D image with DEM rendered images using conditional Generative Adversarial Network architecture (cGAN). We then use this discriminator to search an uncertainty bound given by visual odometry (VO) error bound to estimate rover optimal location and orientation. We demonstrate our network capability to learn to translate rover image into DEM simulated image and match them using Devon Island dataset. The experimental results show that our proposed approach achieves ~74% mean average precision.

• Smart Media Journal
• /
• v.10 no.2
• /
• pp.30-37
• /
• 2021

Recently, 3D contents used in various fields have been attracting people's attention due to the development of virtual reality and augmented reality technology. In order to produce 3D contents, it is necessary to model the objects as vertices. However, high-quality modeling is time-consuming and costly. In order to convert a 2D character into a 3D model, it is necessary to express it as line drawings through feature line extraction. The extraction of consistent line drawings from 2D cartoon cartoons is difficult because the styles and techniques differ depending on the designer who produces them. Therefore, it is necessary to extract the line drawings that show the geometrical characteristics well in 2D cartoon shapes of various styles. This study proposes a method of automatically extracting line drawings. The 2D Cartoon shading image and line drawings are learned by using adversarial network model, which is artificial intelligence technology and outputs 2D cartoon artwork of various styles. Experimental results show the proposed method in this research can be obtained as a result of the line drawings representing the geometric characteristics when a 2D cartoon painting as input.

• Journal of the Korea Convergence Society
• /
• v.12 no.1
• /
• pp.49-55
• /
• 2021

Recently, with the development of database, it is possible to store a lot of data generated in finance, security, and networks. These data are being analyzed through classifiers based on machine learning. The main problem at this time is data imbalance. When we train imbalanced data, it may happen that classification accuracy is degraded due to over-fitting with majority class data. To overcome the problem of data imbalance, oversampling strategy that increases the quantity of data of minority class data is widely used. It requires to tuning process about suitable method and parameters for data distribution. To improve the process, In this study, we propose a strategy to explore and optimize oversampling combinations and ratio based on various methods such as synthetic minority oversampling technique and generative adversarial networks through genetic algorithms. After sampling credit card fraud detection which is a representative case of data imbalance, with the proposed strategy and single oversampling strategies, we compare the performance of trained classifiers with each data. As a result, a strategy that is optimized by exploring for ratio of each method with genetic algorithms was superior to previous strategies.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.1
• /
• pp.57-67
• /
• 2021

Recently, multi-modal deep learning techniques that combine heterogeneous data for deep learning analysis have been utilized a lot. In particular, studies on the synthesis of Text to Image that automatically generate images from text are being actively conducted. Deep learning for image synthesis requires a vast amount of data consisting of pairs of images and text describing the image. Therefore, various data augmentation techniques have been devised to generate a large amount of data from small data. A number of text augmentation techniques based on synonym replacement have been proposed so far. However, these techniques have a common limitation in that there is a possibility of generating a incorrect text from the content of an image when replacing the synonym for a noun word. In this study, we propose a text augmentation method to replace words using word hierarchy information for noun words. Additionally, we performed experiments using MSCOCO data in order to evaluate the performance of the proposed methodology.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.520-523
• /
• 2022

일본 우주항공연구개발기구(Japan Aerospace Exploration Agency, JAXA)는 2007년부터 2017년까지 달 탐사선 셀레네(Selenological and Engineering Explorer, SelEnE)가 관측한 데이터를 수집하고, 연구했다. JAXA는 지구 상층 대기에 존재하는 산소가 자기장의 꼬리 부분에 실려 달로 이동한다는 사실을 발견했다. 하지만 이 연구는 아직 진행 중이며 달의 산화 과정 규명에 추가 연구가 필요하다. 본 논문에서는 생성적 적대 신경망(Generative Adversarial Networks, GAN)으로 달 분화구의 영구 그림자 영역을 제거하고, 물과 얼음을 발견하여 선행 연구의 완성도를 향상하고자 한다. 실험에 사용할 모델은 CIPS(Conditionally Independent Pixel Synthesis)다. CIPS는 실제 같은 영상을 고해상도로 합성한다. 합성할 데이터의 최적인 가중치 초기화 및 파라미터 갱신 방법, 활성 함수 조합은 실험을 통해 확인한다. 필요에 따라 앙상블 학습을 할 수도 있다. 성능평가는 FID(Frechet Inception Distance), 정밀도, 재현율을 사용한다. 제안한 방법은 진행 중인 연구의 시간과 비용을 절약하고, 인과관계를 더욱 명확히 밝히는 데 도움 될 수 있다고 사료된다.

• Journal of Technologic Dentistry
• /
• v.46 no.2
• /
• pp.36-41
• /
• 2024

Purpose: This study aimed to generate virtual mandibular left first molar teeth using deep convolutional generative adversarial networks (DCGANs) and analyze their matching accuracy with actual tooth morphology to propose a new paradigm for using medical data. Methods: Occlusal surface images of the mandibular left first molar scanned using a dental model scanner were analyzed using DCGANs. Overall, 100 training sets comprising 50 original and 50 background-removed images were created, thus generating 1,000 virtual teeth. These virtual teeth were classified based on the number of cusps and occlusal surface ratio, and subsequently, were analyzed for consistency by expert dental technicians over three rounds of examination. Statistical analysis was conducted using IBM SPSS Statistics ver. 23.0 (IBM), including intraclass correlation coefficient for intrarater reliability, one-way ANOVA, and Tukey's post-hoc analysis. Results: Virtual mandibular left first molars exhibited high consistency in the occlusal surface ratio but varied in other criteria. Moreover, consistency was the highest in the occlusal buccal lingual criteria at 91.9%, whereas discrepancies were observed most in the occusal buccal cusp criteria at 85.5%. Significant differences were observed among all groups (p<0.05). Conclusion: Based on the classification of the virtually generated left mandibular first molar according to several criteria, DCGANs can generate virtual data highly similar to real data. Thus, subsequent research in the dental field, including the development of improved neural network structures, is necessary.

• Journal of Wind Energy
• /
• v.13 no.2
• /
• pp.42-52
• /
• 2022

In this paper, we propose a time-series generation methodology using a generative adversarial network (GAN) for long-term prediction of wind and sea conditions, which are information necessary for operations and maintenance (O&M) planning and optimal plans for offshore wind farms. It is a "Conditional TimeGAN" that is able to control time-series data with monthly conditions while maintaining a time dependency between time-series. For the generated time-series data, the similarity of the statistical distribution by direction was confirmed through wave and wind rose diagram visualization. It was also found that the statistical distribution and feature correlation between the real data and the generated time-series data was similar through PCA, t-SNE, and heat map visualization algorithms. The proposed time-series generation methodology can be applied to monthly or annual marine weather prediction including probabilistic correlations between various features (wind speed, wind direction, wave height, wave direction, wave period and their time-series characteristics). It is expected that it will be able to provide an optimal plan for the maintenance and optimization of offshore wind farms based on more accurate long-term predictions of sea and wind conditions by using the proposed model.

• Journal of Platform Technology
• /
• v.10 no.4
• /
• pp.47-55
• /
• 2022

Recently, diverse devices using different wireless technologies are gradually increasing in the IoT environment. In particular, it is essential to design an efficient feature extraction approach and detect the exact types of radio signals in order to accurately identify various radio signal modulation techniques. However, it is difficult to gather labeled wireless signal in a real environment due to the complexity of the process. In addition, various learning techniques based on deep learning have been proposed for wireless signal classification. In the case of deep learning, if the training dataset is not enough, it frequently meets the overfitting problem, which causes performance degradation of wireless signal classification techniques using deep learning models. In this paper, we propose a generative adversarial network(GAN) based on data augmentation techniques to improve classification performance when various wireless signals exist. When there are various types of wireless signals to be classified, if the amount of data representing a specific radio signal is small or unbalanced, the proposed solution is used to increase the amount of data related to the required wireless signal. In order to verify the validity of the proposed data augmentation algorithm, we generated the additional data for the specific wireless signal and implemented a CNN and LSTM-based wireless signal classifier based on the result of balancing. The experimental results show that the classification accuracy of the proposed solution is higher than when the data is unbalanced.