• Journal of Platform Technology
• /
• v.12 no.1
• /
• pp.128-140
• /
• 2024

'Deepfake' refers to a video synthesis technique that utilizes various artificial intelligence technologies to create highly realistic fake content, causing serious confusion to individuals and society by being used for generating fake news, fraud, malicious impersonation, and more. To address this issue, there is a need for methods to detect malicious images generated by deepfake accurately. In this paper, we extract and analyze saliency features from deepfake and real images, and detect candidate synthesis regions on the images, and finally construct an automatic deepfake detection model by focusing on the extracted features. The proposed saliency feature-based model can be universally applied in situations where deepfake detection is required, such as synthesized images and videos. To demonstrate the performance of our approach, we conducted several experiments that have shown the effectiveness of the deepfake detection task.

• Journal of Sensor Science and Technology
• /
• v.30 no.5
• /
• pp.300-308
• /
• 2021

Usually, during slaughter, the meat is divided into large chunks by part after deboning. The meat chunks are inspected for the presence of needles with an X-ray scanner. Although needles in the meat chunks are easily detectable, they can also be found in trimmings and meat offals, where meat skins, fat chunks, and pieces of meat from different parts get agglomerated. Detection of needles in trimmings and meat offals becomes challenging because of many needle-like patterns that are detected by the X-ray scanner. This problem can be solved by learning the trimmings or meat offals using deep learning. However, it is not easy to collect a large number of learning patterns in trimmings or meat offals. In this study, we demonstrate the use of deep convolutional generative adversarial network (DCGAN) to create fake images of trimmings or meat offals and train them using a convolution neural network (CNN).

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.5
• /
• pp.1671-1686
• /
• 2022

Face manipulation tools represented by Deepfake have threatened the security of people's biological identity information. Particularly, manipulation tools with deep learning technology have brought great challenges to Deepfake detection. There are many solutions for Deepfake detection based on traditional machine learning and advanced deep learning. However, those solutions of detectors almost have problems of poor performance when evaluated on different quality datasets. In this paper, for the sake of making high-quality Deepfake datasets, we provide a preprocessing method based on the image pixel matrix feature to eliminate similar images and the residual channel attention network (RCAN) to resize the scale of images. Significantly, we also describe a Deepfake detector named Cascaded-Hop which is based on the PixelHop++ system and the successive subspace learning (SSL) model. By feeding the preprocessed datasets, Cascaded-Hop achieves a good classification result on different manipulation types and multiple quality datasets. According to the experiment on FaceForensics++ and Celeb-DF, the AUC (area under curve) results of our proposed methods are comparable to the state-of-the-art models.

• Communications for Statistical Applications and Methods
• /
• v.30 no.6
• /
• pp.589-603
• /
• 2023

Deep generative models target to infer the underlying true data distribution, and it leads to a huge success in generating fake-but-realistic data. Regarding such a perspective, the data attributes can be a crucial factor in the data generation process since non-existent counterfactual samples can be generated by altering certain factors. For example, we can generate new portrait images by flipping the gender attribute or altering the hair color attributes. This paper proposes counterfactual disentangled variational autoencoder generative adversarial networks (CDVAE-GAN), specialized for data attribute level counterfactual data generation. The structure of the proposed CDVAE-GAN consists of variational autoencoders and generative adversarial networks. Specifically, we adopt a Gaussian variational autoencoder to extract low-dimensional disentangled data features and auxiliary Bernoulli latent variables to model the data attributes separately. Also, we utilize a generative adversarial network to generate data with high fidelity. By enjoying the benefits of the variational autoencoder with the additional Bernoulli latent variables and the generative adversarial network, the proposed CDVAE-GAN can control the data attributes, and it enables producing counterfactual data. Our experimental result on the CelebA dataset qualitatively shows that the generated samples from CDVAE-GAN are realistic. Also, the quantitative results support that the proposed model can produce data that can deceive other machine learning classifiers with the altered data attributes.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.6
• /
• pp.2168-2187
• /
• 2021

The smart grid replaces the traditional power structure with information inventiveness that contributes to a new physical structure. In such a field, malicious information injection can potentially lead to extreme results. Incorrect, FDI attacks will never be identified by typical residual techniques for false data identification. Most of the work on the detection of FDI attacks is based on the linearized power system model DC and does not detect attacks from the AC model. Also, the overwhelming majority of current FDIA recognition approaches focus on FDIA, whilst significant injection location data cannot be achieved. Building on the continuous developments in deep learning, we propose a Deep Learning based Locational Detection technique to continuously recognize the specific areas of FDIA. In the development area solver gap happiness is a False Data Detector (FDD) that incorporates a Convolutional Neural Network (CNN). The FDD is established enough to catch the fake information. As a multi-label classifier, the following CNN is utilized to evaluate the irregularity and cooccurrence dependency of power flow calculations due to the possible attacks. There are no earlier statistical assumptions in the architecture proposed, as they are "model-free." It is also "cost-accommodating" since it does not alter the current FDD framework and it is only several microseconds on a household computer during the identification procedure. We have shown that ANN-MLP, SVM-RBF, and CNN can conduct locational detection under different noise and attack circumstances through broad experience in IEEE 14, 30, 57, and 118 bus systems. Moreover, the multi-name classification method used successfully improves the precision of the present identification.

• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.4
• /
• pp.222-227
• /
• 2022

In this paper, we implemented a Korean text generation and classification model based on a deep learning algorithm that can be applied to various industries. It consists of two implemented GAN-based Korean handwriting generation models and CNN-based Korean handwriting classification models. The GAN model consists of a generator model for generating fake Korean handwriting data and a discriminator model for discriminating fake handwritten data. In the case of the CNN model, the model was trained using the 'PHD08' dataset, and the learning result was 92.45. It was confirmed that Korean handwriting was classified with % accuracy. As a result of evaluating the performance of the classification model by integrating the Korean cursive data generated through the implemented GAN model and the training dataset of the existing CNN model, it was confirmed that the classification performance was 96.86%, which was superior to the existing classification performance.

• Electronics and Telecommunications Trends
• /
• v.35 no.7
• /
• pp.67-76
• /
• 2020

Because of the development and spread of artificial intelligence (AI) technology, new security threats and adverse AI functions have emerged as a real problem in the process of diversifying areas of use and introducing AI-based products and services to users. In response, it is necessary to develop new AI-based technologies in the field of information protection and security. This paper reviews topics such as domestic and international trends on false information detection technology, cyber security technology, and trust distribution platform technology, and it establishes the direction of the promotion of technology development. In addition, the development of international trends in ethical AI guidelines to ensure the human-centered ethical validity of AI development processes and final systems in parallel with technology development are analyzed and discussed. ETRI has developed AI policing technology, information protection, and security technologies as well as derived tasks and implementation strategies to prepare ethical AI development guidelines to ensure the reliability of AI based on its capabilities.

• Annual Conference of KIPS
• /
• 2018.05a
• /
• pp.384-387
• /
• 2018

SNS가 급속도로 확산되며 거짓 정보를 언론으로 위장한 형태인 가짜뉴스는 큰 사회적 문제가 되었다. 본 논문에서는 이를 해결하기 위해 한글 가짜뉴스 탐지를 위한 딥러닝 모델을 제시한다. 기존 연구들은 영어에 적합한 모델들을 제시하고 있으나, 한글은 같은 의미라도 더 짧은 문장으로 표현 가능해 딥러닝을 하기 위한 특징수가 부족하여 깊은 신경망을 운용하기 어렵다는 점과, 형태소 중의성으로 인한 의미 분석의 어려움으로 인해 기존 오델들을 적용하기에는 한계가 있다. 이를 해결하기 위해 얕은 CNN 모델과 음절 단위로 학습된 단어 임베딩 모델인 'Fasttext'를 활용하여 시스템을 구현하고, 이를 학습시켜 검증하였다.

• Journal of Multimedia Information System
• /
• v.9 no.2
• /
• pp.103-112
• /
• 2022

Background: A presentation attack places the printed image or displayed video at the front of the sensor to deceive the biometric recognition system. Usually, presentation attackers steal a genuine user's biometric image and use it for presentation attack. In recent years, reconstruction attack and adversarial attack can generate high-quality fake images, and have high attack success rates. However, their attack rates degrade remarkably after image shooting. Methods: In order to comprehensively analyze the threat of presentation attack to palmprint recognition system, this paper makes six palmprint presentation attack datasets. The datasets were tested on texture coding-based recognition methods and deep learning-based recognition methods. Results and conclusion: The experimental results show that the presentation attack caused by the leakage of the original image has a high success rate and a great threat; while the success rates of reconstruction attack and adversarial attack decrease significantly.

• Annual Conference of KIPS
• /
• 2021.05a
• /
• pp.340-342
• /
• 2021

흔히 DeepFake로 불리는 GAN 기술은 소스 영상과 타겟 이미지를 합성하여 타겟 이미지 내의 사람이 소스 영상에서 나타나도록 합성하는 기술이다. 이러한 GAN 기반 영상 합성 기술은 2018년을 기점으로 급격한 성장세를 보이며 다양한 산업에 접목되어지고 있으나 학습 모델을 얻는 데 걸리는 시간이 너무 오래 소요되고, 감정 표현을 인지하는 데 어려움이 있었다. 본 논문에서는 상기 두가지 문제를 해결하기 위해 Facial Action Coding System(FACS) 및 음성 합성 기술[4]을 적용한 가상 아바타 생성 방법에 대해 제안하고자 한다.