Trans-DF: A Transfer Learning-based end-to-end Deepfake Detector

Patel, Mohil Maheshkumar; Gupta, Aaryan; Tanwar, Sudeep; Obaidat, Mohammad S.

doi:10.1109/iccca49541.2020.9250803

Cited by 14 publications

(12 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also, the image size of the dataset was only 128 × 128 px, whereas the method mentioned above uses a 224 × 224 px image. In the research paper [27], a transfer learning method based on the GAN architecture was used to detect deepfake images. A few methods of transfer learning, like VGG16 and MobileNet, were carried out in the paper [27], but every method gave average accuracy.…”

Section: Model Comparisonmentioning

confidence: 99%

“…In the research paper [27], a transfer learning method based on the GAN architecture was used to detect deepfake images. A few methods of transfer learning, like VGG16 and MobileNet, were carried out in the paper [27], but every method gave average accuracy. The method mentioned in this paper has low validation loss and high validation accuracy.…”

Section: Model Comparisonmentioning

confidence: 99%

“…Not only for deepfake images, but some researchers have also worked to detect deepfake videos and audio. One of the approaches [27] proposes the idea of detecting and extracting human faces from videos and then classifying them as real or fake. They used various CNN methods, such as VGG, ResNet, and Inception, to detect deepfakes.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Detecting Deepfake Images Using Deep Learning Techniques and Explainable AI Methods

Abir¹,

Khanam²,

Alam³

et al. 2023

Intelligent Automation &Amp; Soft Computing

View full text Add to dashboard Cite

Section: Model Comparisonmentioning

confidence: 99%

Section: Model Comparisonmentioning

confidence: 99%

See 1 more Smart Citation

Detecting Deepfake Images Using Deep Learning Techniques and Explainable AI Methods

Abir¹,

Khanam²,

Alam³

et al. 2023

Intelligent Automation &Amp; Soft Computing

View full text Add to dashboard Cite

“…We discovered in a study that the graph of photo-shoped faked and morphed video climbs dramatically in various social sites such as in digital marking, political marketing, etc which helps the society in a very positive manner and helps them in very different approaches using artificial intelligence. According to the study done by sensity.ai published shows that the content of Deep_fake media is just double every six months as shown in figure 1 till 2020 [5]. This report by the World Economic Forum [6.] on the publication of research papers on Deep_fake, demonstrates that it's a very hot topic not only in the internet world but also in the research world.…”

Section: Introductionmentioning

confidence: 99%

Analysis of Deep-Fake Technology Impacting Digital World Credibility: A Comprehensive Literature Review

Akbar

Suaib

Hussain³

2022

IJCIT

View full text Add to dashboard Cite

Deep-Fake Technique is a new scientific method that uses Artificial-Intelligince to make fake videos with an affect of facial expressions and coordinated movement of lips. This technology is frequently employed in a variety of contexts with various goals. Deep-Fake technology is being used to generate an extremely realistic fake video that can be widely distributed to promote false information or fake news about any celebrity or leader that was not created by them. Because of the widespread use of social media, these fraudulent videos can garner billions of views in under an hour and have a significant impact on our culture. Deep-Fakes are a threat to our celebrities, democracy, religious views, and commerce, according to the findings, but they can be managed through rules and regulations, strong company policy, and general internet user awareness and education. We need to devise a process for examining such video and distinguishing between actual and fraudulent footage.

show abstract

“…Researchers have also proposed different deepfake datasets generated using the latest deepfake generation methods [33][34][35][36][37][38] to aid other fellow scholars in developing deepfake detection methods. Most of the recent deepfake detection methods [18,22,23,26,30,[30][31][32]39] make use of these publicly available datasets. However, these deepfake datasets only focus on generating realistic deepfake videos but do not consider generating fake audio for them.…”

Section: Introductionmentioning

confidence: 99%

FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Dataset

Khalid¹,

Tariq²,

Kim³

et al. 2021

Preprint

View full text Add to dashboard Cite

While the significant advancements have made in the generation of deepfakes using deep learning technologies, its misuse is a well-known issue now. Deepfakes can cause severe security and privacy issues as they can be used to impersonate a person's identity in a video by replacing his/her face with another person's face. Recently, a new problem of generating synthesized human voice of a person is emerging, where AI-based deep learning models can synthesize any person's voice requiring just a few seconds of audio. With the emerging threat of impersonation attacks using deepfake audios and videos, a new generation of deepfake detectors is needed to focus on both video and audio collectively. A large amount of good quality dataset is typically required to capture the real-world scenarios to develop a competent deepfake detector. Existing deepfake datasets either contain deepfake videos or audios, which are racially biased as well. Hence, there is a crucial need for creating a good video as well as audio deepfake dataset, which can be used to detect audio and video deepfake simultaneously. To fill this gap, we propose a novel Audio-Video Deepfake dataset (FakeAVCeleb) that contains not only deepfake videos but also respective synthesized lip-synced fake audios. We generate this dataset using the current most popular deepfake generation methods. We selected real YouTube videos of celebrities with four racial backgrounds (Caucasian, Black, East Asian, and South Asian) to develop a more realistic multimodal dataset that addresses racial bias, and further help develop multimodal deepfake detectors. We performed several experiments using state-of-the-art detection methods to evaluate our deepfake dataset and demonstrate the challenges and usefulness of our multimodal Audio-Video deepfake dataset.Preprint. Under review.

show abstract

Trans-DF: A Transfer Learning-based end-to-end Deepfake Detector

Cited by 14 publications

References 17 publications

Detecting Deepfake Images Using Deep Learning Techniques and Explainable AI Methods

Detecting Deepfake Images Using Deep Learning Techniques and Explainable AI Methods

Analysis of Deep-Fake Technology Impacting Digital World Credibility: A Comprehensive Literature Review

FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Dataset

Contact Info

Product

Resources

About