• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.39-60
• /
• 2021

In the field of deep learning, there are many algorithms mainly after GAN in research related to generation, but in terms of generation, there are similarities and differences with art. If the generation in the engineering aspect is mainly to judge the presence or absence of a quantitative indicator or the correct answer and the incorrect answer, the creation in the artistic aspect creates a creation that interprets the world and human life by cross-validating and doubting the correct answer and incorrect answer from various perspectives. In this paper, the video generation ability of deep learning was interpreted from the perspective of collage and compared with the results made by the artist. The characteristic of the experiment is to compare and analyze how much GAN reproduces the result of the creator made with the collage technique and the difference between the creative part, and investigate the satisfaction level by making performance evaluation items for the reproducibility of GAN. In order to experiment on how much the creator's statement and purpose of expression were reproduced, a deep learning algorithm corresponding to the statement keyword was found and its similarity was compared. As a result of the experiment, GAN did not meet much expectations to express the collage technique. Nevertheless, the image association showed higher satisfaction than human ability, which is a positive discovery that GAN can show comparable ability to humans in terms of abstract creation.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.7
• /
• pp.303-314
• /
• 2023

Voice conversion, a technology that allows an individual's speech data to be regenerated with the acoustic properties(tone, cadence, gender) of another, has countless applications in education, communication, and entertainment. This paper proposes an approach based on the StarGAN-VC model that generates realistic-sounding speech without requiring parallel utterances. To overcome the constraints of the existing StarGAN-VC model that utilizes one-hot vectors of original and target speaker information, this paper extracts feature vectors of target speakers using a pre-trained version of Rawnet3. This results in a latent space where voice conversion can be performed without direct speaker-to-speaker mappings, enabling an any-to-any structure. In addition to the loss terms used in the original StarGAN-VC model, Wasserstein distance is used as a loss term to ensure that generated voice segments match the acoustic properties of the target voice. Two Time-Scale Update Rule (TTUR) is also used to facilitate stable training. Experimental results show that the proposed method outperforms previous methods, including the StarGAN-VC network on which it was based.

• Journal of Internet Computing and Services
• /
• v.21 no.5
• /
• pp.129-137
• /
• 2020

Recently, software size and program code complexity have increased due to application to various fields of software. Accordingly, the existence of program bugs inevitably occurs, and the cost of software maintenance is increasing. In open source projects, developers spend a lot of debugging time when solving a bug report assigned. To solve this problem, in this paper, we apply SeqGAN algorithm to software bug repair. In detail, the SeqGAN model is trained based on the source code. Open similar source codes during the learning process are also used. To evaluate the suitability for the generated candidate patch, a fitness function is applied, and if all test cases are passed, software bug correction is considered successful. To evaluate the efficiency of the proposed model, it was compared with the baseline, and the proposed model showed better repair.

• Smart Media Journal
• /
• v.11 no.1
• /
• pp.38-45
• /
• 2022

A lot of research is being going until this day in order to obtain stable profit in the stock market. Trading algorithms are widely used, accounting for over 80% of the trading volume of the US stock market. Despite a lot of research, there is no trading algorithm that always shows good performance. In other words, there is no guarantee that an algorithm that performed well in the past will perform well in the future. The reason is that there are many factors that affect the stock price and there are uncertainties about the future. Therefore, in this paper, we propose a model using TimeGAN that predicts future returns well and selects algorithms that are expected to have high returns based on past records of the returns of algorithms. We use TimeGAN becasue it is probabilistic, whereas LSTM method predicts future time series data is deterministic. The advantage of TimeGAN probabilistic prediction is that it can reflect uncertainty about the future. As an experimental result, the method proposed in this paper achieves a high return with little volatility and shows superior results compared to many comparison algorithms.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.06a
• /
• pp.36-39
• /
• 2017

2014 년 Ian Goodfellow 가 발표한 한편의 논문은 머신러닝 분야에 새로운 방향을 제시하였다. Generative Adversarial Networks, 일명 GAN 이라 불리는 이 논문은 이전까지 딥러닝으로 하지못했던 새로운 것을 창조해내는 작업을 하는 첫번째 딥러닝 알고리즘이다. 이전까지는 딥러닝을 통해 영상에서 객체의 종류를 판단하는 Classification 문제나, 영상에서 특정 객체를 검출하여 위치를 찾는 Object detection, 영상 내 특정 객체만 분리해내는 Image segmentation 문제를 해결하고 있었다. GAN 의 등장으로, 다양한 방면에서 GAN 을 적용하여 기존에는 하지 못했던 새로운 분야에 딥러닝을 적용한 사례들이 등장하고 있다. 본 논문에서는 GAN 의 원리 분석과 GAN 을 응용하여 여러 분야에 적용한 사례들을 살펴보고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.11a
• /
• pp.792-794
• /
• 2017

현재 머신러닝(Machine Learning) 기술은 기존의 머신러닝과 조합 및 변형 되어 조금 더 발전 된 형태로 연구되어지고 있다. 따라서 수많은 알고리즘이 개발되고 있는 시점이다. 본 연구는 최근 좋은 결과로 관심을 받고있는 GAN(Generative Adversarial Net)을 중심으로 IT기술의 머신러닝과 그림을 조합하여 자동채색을 목적으로 GAN 알고리즘을 비교하고 분석하고자 한다. GAN 알고리즘들 가운데서 'Conditional GAN'과 'Wasserstein GAN'을 사용하여 자동채색을 적용시켰고, 가장 부합한 알고리즘을 찾고 성능을 비교하여 어떠한 알고리즘이 '자동채색' 목적에 더 부합한지 비교하고 판단 한다.

• Journal of Korea Multimedia Society
• /
• v.23 no.12
• /
• pp.1476-1485
• /
• 2020

Segmenting OCT retinal images into layers is important to diagnose and understand the progression of retinal diseases or identify potential symptoms. The task of manually identifying these layers is a difficult task that requires a lot of time and effort even for medical professionals, and therefore, various studies are being conducted to automate this using deep learning technologies. In this paper, we use cGAN-based neural network to automatically segmenting OCT retinal images into seven terrain-type regions defined by six layer boundaries. The network is composed of a Segnet-based generator model and a discriminator model. We also proposed a dynamic programming algorithm for refining the outputs of the network. We performed experiments using public OCT image data set and compared its performance with the Segnet-only version of the network. The experimental results show that the cGAN-based network outperforms Segnet-only version.

• Journal of the Korean Society of Visualization
• /
• v.19 no.3
• /
• pp.63-68
• /
• 2021

In this study, data was augmentation through the SinGAN algorithm using small image data, and defects in rubber O-rings were detected using the random forest algorithm. Unlike the commonly used data augmentation image rotation method to solve the data imbalance problem, the data imbalance problem was solved by using the SinGAN algorithm. A study was conducted to distinguish between normal products and defective products of rubber o-ring by using the random forest algorithm. A total of 20,000 image date were divided into transit and testing datasets, and an accuracy result was obtained to distinguish 97.43% defects as a result of the test.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.530-531
• /
• 2020

최근 인공지능을 활용하여 예술 작품에 몰입할 수 있도록 무대 효과를 디자인하는 연구가 진행되고 있다. 무대 효과 중에서 무대 배경은 공연의 분위기를 형성한다. 춤의 장르별로 무대 배경에 사용되는 이미지를 생성하기 위해 소셜 미디어 기반 무대 배경 생성 시스템이 있다. 하지만 같은 장르 춤은 동일한 무대 배경 이미지가 제공되는 문제가 있다. 같은 장르의 춤이지만 노래의 분위기를 반영하여 차별된 무대 배경 이미지를 제공하는 것이 필요하다. 본 논문은 노래 가사의 감정을 활용하여 Generative Adversarial Network(GAN)을 통해 각 노래의 분위기를 고려한 무대 배경 이미지를 생성하는 방법을 제안한다. GAN은 노래에 포함된 단락별 감정 단어를 추출하여 스타일을 생성하도록 학습된다. 학습된 GAN은 노래 가사에 포함된 감정 단어를 활용하여 곡의 분위기를 반영한 무대 배경 이미지를 생성한다. 노래 가사를 고려하여 무대 배경 이미지를 생성함으로써 곡의 분위기가 고려된 무대 배경 이미지 생성이 가능하다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.124-125
• /
• 2023

Defect detection during liquid crystal display (LCD) manufacturing has always been a critical challenge. This study aims to address this issue by proposing a data augmentation method based on generative adversarial networks (GAN) to improve defect identification accuracy in LCD production. By leveraging synthetically generated image data from GAN, we effectively augment the original dataset to make it more representative and diverse. This data augmentation strategy enhances the model's generalization capability and robustness on real-world data. Compared to traditional data augmentation techniques, the synthetic data from GAN are more realistic, diverse and broadly distributed. Experimental results demonstrate that training models with GAN-generated data combined with the original dataset significantly improves the detection accuracy of critical defects in LCD manufacturing, compared to using the original dataset alone. This study provides an effective data augmentation approach for intelligent quality control in LCD production.