• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.6
• /
• pp.1053-1065
• /
• 2020

With the recent development of hardware performance and artificial intelligence technology, sophisticated fake videos that are difficult to distinguish with the human's eye are increasing. Face synthesis technology using artificial intelligence is called Deepfake, and anyone with a little programming skill and deep learning knowledge can produce sophisticated fake videos using Deepfake. A number of indiscriminate fake videos has been increased significantly, which may lead to problems such as privacy violations, fake news and fraud. Therefore, it is necessary to detect fake video clips that cannot be discriminated by a human eyes. Thus, in this paper, we propose a deep-fake detection model applied with Bidirectional Convolution LSTM and Attention Module. Unlike LSTM, which considers only the forward sequential procedure, the model proposed in this paper uses the reverse order procedure. The Attention Module is used with a Convolutional neural network model to use the characteristics of each frame for extraction. Experiments have shown that the model proposed has 93.5% accuracy and AUC is up to 50% higher than the results of pre-existing studies.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.8
• /
• pp.31-38
• /
• 2023

Style transfer is one of deep learning-based image processing techniques that has been actively researched recently. These research efforts have led to significant improvements in the quality of result images. Style transfer is a technology that takes a content image and a style image as inputs and generates a transformed result image by applying the characteristics of the style image to the content image. It is becoming increasingly important in exploiting the diversity of digital content. To improve the usability of style transfer technology, ensuring stable performance is crucial. Recently, in the field of natural language processing, the concept of Transformers has been actively utilized. Attention maps, which forms the basis of Transformers, is also being actively applied and researched in the development of style transfer techniques. In this paper, we analyze the representative techniques SANet and AdaAttN and propose a novel attention map-based structure which can generate improved style transfer results. The results demonstrate that the proposed technique effectively preserves the structure of the content image while applying the characteristics of the style image.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.4
• /
• pp.345-356
• /
• 2023

Target signal of passive sonar shows narrow band harmonic characteristic with a variation in intensity within a few seconds and long term frequency variation due to the Lloyd's mirror effect. We propose a signal classification algorithm based on Gated Recurrent Unit (GRU) that learns local and global time series features. The algorithm proposed implements a multi layer network using GRU and extracts local and global time series features via dilated connections. We learns attention mechanism to weight time series features and classify passive sonar signals. In experiments using public underwater acoustic data, the proposed network showed superior classification accuracy of 96.50 %. This result is 4.17 % higher classification accuracy compared to existing skip connected GRU network.

• Smart Media Journal
• /
• v.12 no.11
• /
• pp.103-112
• /
• 2023

Recently, the need to prevent battery fires and accidents has emerged, as the use of lithium-ion batteries has increased. In order to prevent accidents, it is necessary to predict the state of health (SOH) and check the replacement timing of the battery with a lot of degradation. This paper proposes a model for predicting the degradation state of a battery by using four battery degradation indicators: maximum voltage arrival time, current change time, maximum temperature arrival time, and incremental capacity (IC) that can be obtained in the battery charging process, and LSTM using an attention mechanism. The performance of the proposed model was measured using the NASA battery data set, and the predictive performance was improved compared to that of the general LSTM model, especially in the SOH 90-70% section, which is close to the battery replacement cycle.

• Journal of the Korea Computer Graphics Society
• /
• v.28 no.3
• /
• pp.55-65
• /
• 2022

Recently, deep learning-based automated systems for identifying and detecting landmarks have been proposed. In order to train such a deep learning-based model without overfitting, a large amount of image and labeling data is required. Conventionally, an experienced reader manually identifies and labels landmarks in a patient's image. However, such measurement is not only expensive, but also has poor reproducibility, so the need for an automated labeling method has been raised. In addition, in the X-ray image, since various human tissues on the path through which the photons pass are displayed, it is difficult to identify the landmark compared to a general natural image or a 3D image modality image. In this study, we propose a geometric data augmentation technique that enables the generation of a large amount of labeling data in X-ray images. In addition, the optimal attention mechanism for landmark detection was presented through the implementation and application of various attention techniques to improve the detection performance of 16 major landmarks in the skull. Finally, among the major cranial landmarks, markers that ensure stable detection are derived, and these markers are expected to have high clinical application potential.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.270-273
• /
• 2021

한국어 의존 파싱은 전이 기반 방식과 그래프 기반 방식의 두 갈래로 연구되어 왔고 그 중 그래프 기반 의존 파싱 방법은 문장 내의 모든 단어에 대해 인코딩한 후 지배소, 의존소에 대한 MLP를 적용하여 각각 표상을 얻고 Biaffine 어텐션을 통해 모든 단어 쌍에 대한 그래프 점수를 얻고 트리를 생성하는 방법이 대표적이다. Biaffine 어텐션 모델에서 문장 내의 각 단어들은 구문 트리 내의 서브트리의 역할을 하지만 두 단어간의 의존성만을 판단하기 때문에 서브 트리의 정보를 이용할 수 없다는 단점이 존재한다. 본 연구에서는 이러한 단점을 해결하기 위해 제안된 Span-Span(서브트리-서브트리)로의 서브트리 정보를 이용할 수 있도록 하는 기계 독해 기반 의존 파싱 모델을 한국어 구문 분석 데이터 셋에 적용하여 소폭의 성능향상을 얻었다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.12
• /
• pp.1880-1889
• /
• 2022

The online portal platform provides online news with online comments, but the anonymity of comments causes incivility, and online comments are considered social problems. While there are many foreign language-based incivility detection studies, in-depth research is not being conducted in Korea since there has not been implemented Korean language dataset which is labeled detailed criteria of incivility. In this study, the incivility notation of comments was conducted in a total of 13 items, uncivil words were summarized. Furthermore, Attention algorithm was applied to each comment and summary to extract embedding vectors. 2-d CNN followed at the end to detect incivility in given data. As a result, we showed that the proposed algorithm is useful for anti-citizen detection such as name-calling and offensive tones. This study is expected to contribute to the formation of a healthy online comment culture by detecting uncivil comments which hinder democratic discourse.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.6
• /
• pp.544-551
• /
• 2023

Speech enhancement used to improve the perceptual quality and intelligibility of noise speech has been studied as a method using a complex-valued spectrum that can improve both magnitude and phase in a method using a magnitude spectrum. In this paper, a study was conducted on how to apply attention mechanism to complex-valued spectrum-based speech enhancement systems to further improve the intelligibility and quality of noise speech. The attention is performed based on additive attention and allows the attention weight to be calculated in consideration of the complex-valued spectrum. In addition, the global average pooling was used to consider the importance of the feature map. Complex-valued spectrum-based speech enhancement was performed based on the Deep Complex U-Net (DCUNET) model, and additive attention was conducted based on the proposed method in the Attention U-Net model. The results of the experiments on noise speech in a living room environment showed that the proposed method is improved performance over the baseline model according to evaluation metrics such as Source to Distortion Ratio (SDR), Perceptual Evaluation of Speech Quality (PESQ), and Short Time Object Intelligence (STOI), and consistently improved performance across various background noise environments and low Signal-to-Noise Ratio (SNR) conditions. Through this, the proposed speech enhancement system demonstrated its effectiveness in improving the intelligibility and quality of noisy speech.

• The Journal of Korean Association of Computer Education
• /
• v.12 no.6
• /
• pp.21-32
• /
• 2009

This paper proposes a method to extract a user's intention automatically and implementation of intention map that support a user can appropriate search results using a user' information need accurately. It selects user intention based on searching history obtained from previous users' same queries and extracts user intentions by using clustering algorithm and user intention extraction algorithm, extracted user intentions are represented in an intention map base on a theory of knowledge representation. For the efficiency analysis of intention map, we extracted user intentions using 2,600 search history data which provided by a current domestic commercial search engine. The experimental results using the information intention map search when using general search engines represent more than satisfaction was statistically significant.

• The Korean Journal of Applied Statistics
• /
• v.34 no.2
• /
• pp.191-204
• /
• 2021

Visual Question Answering (VQA) and image captioning are tasks that require understanding of the features of images and linguistic features of text. Therefore, co-attention may be the key to both tasks, which can connect image and text. In this paper, we propose a model to achieve high performance for VQA by image caption generated using a pretrained standard transformer model based on MSCOCO dataset. Captions unrelated to the question can rather interfere with answering, so some captions similar to the question were selected to use based on a similarity to the question. In addition, stopwords in the caption could not affect or interfere with answering, so the experiment was conducted after removing stopwords. Experiments were conducted on VQA-v2 data to compare the proposed model with the deep modular co-attention network (MCAN) model, which showed good performance by using co-attention between images and text. As a result, the proposed model outperformed the MCAN model.