• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.11a
• /
• pp.172-175
• /
• 2007

영상에 나타나는 자막은 영상과 관계가 있는 정보를 포함한다. 이러한 자막의 정보를 이용하기 위해 영상으로부터 자막을 추출하는 연구는 근래에 들어 활발히 진행되고 있다. 기존의 연구는 일정한 높이의 자막이나 획의 두께를 가진 자막만을 추출하였다. 본 논문에서는 일정 크기 이상의 크기에 무관한 자막을 추출하는 방법을 제안한다. 먼저, 자막 추출을 위해서 영상에 포함된 픽셀들의 연결 객체를 생성하였다. 그리고 연결 객체 중에서 자막의 형태적인 특정의 패턴을 분석하고, 패턴을 이용하여 자막을 추출하였다. 실험에 사용된 영상은 다큐멘터리, 쇼 프로그램과 같은 대중 방송에서 획득하였으며, 실험 결과는 다양한 크기의 자막을 포함한 영상을 사용하여 실험하였고, 자막 추출의 결과는 찾아진 연결객체 중에 자막의 비율과 자막 중에 찾아진 자막의 비율로 분석하였다. 제안한 방법에 의해 다양한 크기의 자막을 추출할 수 있었다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.11a
• /
• pp.66-69
• /
• 2017

본 논문에서는 깊은 합성 곱 신경망 (Deep Convolutional Neural Network) 를 이용해서 SAR (Synthetic Aperture Radar) 영상의 반전 잡음 (speckle noise) 성분을 제거하는 기법을 제안하고자 한다. Deep Convolutional Neural Network는 이미지의 데이터 특성에 적합한 딥 러닝 방법이고, 이는 SAR 위성영상의 반전 잡음 제거에 사용해도 효과적이다. 반전 잡음 필터 모델 추정을 위한 학습은 임의로 반전 잡음을 합성한 트레이닝 이미지들과 원본 트레이닝 이미지들을 이용한 회귀모델을 통해 진행된다. 학습을 통해 얻은 반전 잡음 필터는 기존 알고리즘에 비해 우수한 외곽선 보존 성능을 나타냄을 확인하였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.176-179
• /
• 2021

최근 딥 러닝 기법의 하나인 합성곱 신경망(Convolutional Neural Network, CNN)은 영상 잡음(Noise) 제거 분야에서 전통적인 기법보다 좋은 성능을 나타내고 있지만 학습하는 과정에서 영상 내 디테일한 부분이 손실될 수 있다. 본 논문에서는 웨이블릿 변환(Wavelet Transform)을 기반으로 영상 내 디테일 정보도 같이 학습하여 영상 디테일을 향상하는 잡음 제거 합성곱 신경망 네트워크를 제안한다. 제안하는 네트워크는 디테일 향상 서브 네트워크(Detail Enhancement Subnetwork)와 영상 잡음 추출 서브 네트워크(Noise Extraction Subnetwork)를 이용하게 된다. 실험을 통해 제안하는 방법은 기존 알고리듬보다 디테일 손실 문제를 효과적으로 해결할 수 있었고 객관적 품질 평가인 PSNR(Peak Signal-to-Noise Ratio)와 주관적 품질 비교에서 모두 우수한 결과가 나온 것을 확인하였다.

• Journal of Broadcast Engineering
• /
• v.24 no.2
• /
• pp.217-226
• /
• 2019

In this paper, we propose a filtering method that can improve the results of inverse tone-mapping using guided image filter. Inverse tone-mapping techniques have been proposed that convert LDR images to HDR. Recently, many algorithms have been studied to convert single LDR images into HDR images using CNN. Among them, there exists an algorithm for restoring pixel information using CNN which learned to restore saturated region. The algorithm does not suppress the noise in the non-saturation region and cannot restore the detail in the saturated region. The proposed algorithm suppresses the noise in the non-saturated region and restores the detail of the saturated region using a WGIF in the input image, and then applies it to the CNN to improve the quality of the final image. The proposed algorithm shows a higher quantitative image quality index than the existing algorithms when the HDR quantitative image quality index was measured.

• Journal of the Korea Computer Graphics Society
• /
• v.27 no.4
• /
• pp.11-24
• /
• 2021

In this paper, we propose an artificial neural network framework that can represent the foam effects expressed in liquid simulation in detail without noise. The position and advection of foam particles are calculated using the existing screen projection method, and the noise problem that appears in this process is solved through an proposed artificial neural network. The important thing in the screen projection approach is the projection map, but noise occurs in the projection map in the process of projecting momentum into the discretized screen space, and we efficiently solve this problem by using an artificial neural network-based denoising network. When the foam generating area is selected through the projection map, 2D is inversely transformed into 3D space to generate foam particles. We solve the existing denoising network problem in which small-scaled foam particles disappear. In addition, by integrating the proposed algorithm with the screen-space projection framework, all the advantages of this approach can be accommodated. As a result, it shows through various experiments whether it is possible to stably represent not only the clean foam effects but also the foam particles lost due to the denoising process.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.3
• /
• pp.115-122
• /
• 2019

For Korean phoneme recognition, Hidden Markov-Gaussian Mixture model(HMM-GMM) or hybrid models which combine artificial neural network with HMM have been mainly used. However, current approach has limitations in that such models require force-aligned corpus training data that is manually annotated by experts. Recently, researchers used neural network based phoneme recognition model which combines recurrent neural network(RNN)-based structure with connectionist temporal classification(CTC) algorithm to overcome the problem of obtaining manually annotated training data. Yet, in terms of implementation, these RNN-based models have another difficulty in that the amount of data gets larger as the structure gets more sophisticated. This problem of large data size is particularly problematic in the Korean language, which lacks refined corpora. In this study, we introduce CTC algorithm that does not require force-alignment to create a Korean phoneme recognition model. Specifically, the phoneme recognition model is based on convolutional neural network(CNN) which requires relatively small amount of data and can be trained faster when compared to RNN based models. We present the results from two different experiments and a resulting best performing phoneme recognition model which distinguishes 49 Korean phonemes. The best performing phoneme recognition model combines CNN with 3hop Bidirectional LSTM with the final Phoneme Error Rate(PER) at 3.26. The PER is a considerable improvement compared to existing Korean phoneme recognition models that report PER ranging from 10 to 12.

• Journal of Korea Multimedia Society
• /
• v.11 no.9
• /
• pp.1213-1226
• /
• 2008

A recommender system supports people in making recommendations finding a set of people who are likely to provide good recommendations for a given person, or deriving recommendations from implicit behavior such as browsing activity, buying patterns, and time on task. We proposed new recommender system which combined SOM(Self-Organizing Map) neural networks with the Collaborative filtering which most recommender systems hat applied First, we segmented user groups according to demographic characteristics and then we trained the SOM with people's preferences as ito inputs. Finally we applied the classic collaborative filtering to the clustering with similarity in which an recommendation seeker belonged to, and therefore we didn't have to apply the collaborative filtering to the whose data set. Experiments were run for EachMovies data set. The results indicated that the predictive accuracy was increased in terms of MAE(Mean-Absolute-Error).

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10b
• /
• pp.295-297
• /
• 2004

생체 내의 모든 기능은 유전자 발현에 의해 결정된다. 유전자 발현은 않은 인자들에 의해 조절되며, 이러한 조절 과정에 따라 유전자 발현량이 결정되는 것이다. 세포 주기 역시 유전자 발현과 밀접한 연관성을 가지고 있다. 본 논문에서는 효모에서 세포 주기의 각 단계와 관련된 유전자들의 분석을 통해서 세포주기를 조절하는데 있어서 중요한 역할을 수행하는 전사 조절 모티프들이 무엇인지를 찾아보았다. 주요 모티프의 추출은 인공신경망 모델을 학습하고. 입출력 에러 분석을 통하여 이루어진다. 그 결과 MCB 등 기존의 실험 결과를 통하여 세포주기에 관련이 있다고 알려진 모티프들이 높은 점수를 보인다는 것을 알 수 있었고. 그 외에 세포주기의 각 단계에서 유전자 발현에 중요한 역할을 수행할 것으로 예상되는 다른 모티프들도 예측해볼 수 있었다.

• Journal of Broadcast Engineering
• /
• v.23 no.4
• /
• pp.484-494
• /
• 2018

The uniform motion blur removing algorithm of Cho et al. has the problem that the edge region of the image cannot be restored clearly. We propose the effective algorithm to overcome this problem by using shock filter that reconstructs a blurred step signal into a sharp edge, and convolutional neural network (CNN) that learns by extracting features from the image. Then uniform motion blur kernel is estimated from the latent sharp image to remove blur in the image. The proposed algorithm improved the disadvantages of the conventional algorithm by reconstructing the latent sharp image using shock filter and CNN. Through the experimental results, it was confirmed that the proposed algorithm shows excellent reconstruction performance in objective and subjective image quality than the conventional algorithm.

• Journal of Digital Contents Society
• /
• v.19 no.3
• /
• pp.435-443
• /
• 2018

Recently, with the explosive development of computing power, various methods such as RNN and CNN have been proposed under the name of Deep Learning, which solve many problems of Computer Vision have. The Generative Adversarial Network, released in 2014, showed that the problem of computer vision can be sufficiently solved in unsupervised learning, and the generation domain can also be studied using learned generators. GAN is being developed in various forms in combination with various models. Machine learning has difficulty in collecting data. If it is too large, it is difficult to refine the effective data set by removing the noise. If it is too small, the small difference becomes too big noise, and learning is not easy. In this paper, we apply a deep CNN model for extracting facial region in image frame to GAN model as a preprocessing filter, and propose a method to produce composite images of various facial expressions by stably learning with limited collection data of two persons.