A YOLOV3-Based Method for Detecting Deepfake Manipulated Facial Images

Mehmet Karaköse; Hasan Yetiş; Mert Çeçen

Makale detayları

A YOLOV3-Based Method for Detecting Deepfake Manipulated Facial Images

Yazarlar:

Mehmet Karaköse | Hasan Yetiş | Mert Çeçen

Yayıncı:

Fırat University

Ana Başlıklar:

Deep Learning

Dergi:

Turkish Journal of Science and Technology Vol.19 Issue 2, pp.315-324

Belge Tipi:

Makale

Yayın Tarihi:

9/2024

Belge Dili:

İngilizce

Document ID (Org No.):

2001328531

pISSN:

1308-9080

Anahtar kelimeler:

Özet:

 With the advancement of technology and the development of applications that make it easier to transfer images, sounds and videos to the virtual environment, it has become much easier to access peoples personal information, videos and images. Deepfake technology produces fakes of authentic images or sounds using deep learning and artificial intelligence techniques. Today, in addition to being used in the entertainment and film industries, it is also used in situations such as creating fake news and discrediting people. Different studies have been conducted in the literature to detect deepfake images and videos to prevent these situations. In this study, a comprehensive literature review was conducted. Real and fake images were collected and labelled from different datasets or videos, and a dataset was created by applying the necessary pre-processing steps. With the created dataset, training was carried out with YOLOv3 technology, which calculates class probabilities differently from traditional methods using Convolutional Neural Networks CNN and handles all operations in a single regression problem, which can make fast and high-accurate detection, and the modelling process is explained. With the tests performed in the study, the model that can detect fake images produced with deepfake technology with 95 accuracy was obtained.

 With the advancement of technology and the development of applications that make it easier to transfer images, sounds and videos to the virtual environment, it has become much easier to access peoples personal information, videos and images. Deepfake technology produces fakes of authentic images or sounds using deep learning and artificial intelligence techniques. Today, in addition to being used in the entertainment and film industries, it is also used in situations such as creating fake news and discrediting people. Different studies have been conducted in the literature to detect deepfake images and videos to prevent these situations. In this study, a comprehensive literature review was conducted. Real and fake images were collected and labelled from different datasets or videos, and a dataset was created by applying the necessary pre-processing steps. With the created dataset, training was carried out with YOLOv3 technology, which calculates class probabilities differently from traditional methods using Convolutional Neural Networks CNN and handles all operations in a single regression problem, which can make fast and high-accurate detection, and the modelling process is explained. With the tests performed in the study, the model that can detect fake images produced with deepfake technology with 95 accuracy was obtained.

 With the advancement of technology and the development of applications that make it easier to transfer images, sounds and videos to the virtual environment, it has become much easier to access peoples personal information, videos and images. Deepfake technology produces fakes of authentic images or sounds using deep learning and artificial intelligence techniques. Today, in addition to being used in the entertainment and film industries, it is also used in situations such as creating fake news and discrediting people. Different studies have been conducted in the literature to detect deepfake images and videos to prevent these situations. In this study, a comprehensive literature review was conducted. Real and fake images were collected and labelled from different datasets or videos, and a dataset was created by applying the necessary pre-processing steps. With the created dataset, training was carried out with YOLOv3 technology, which calculates class probabilities differently from traditional methods using Convolutional Neural Networks CNN and handles all operations in a single regression problem, which can make fast and high-accurate detection, and the modelling process is explained. With the tests performed in the study, the model that can detect fake images produced with deepfake technology with 95 accuracy was obtained.

 Teknolojinin ilerlemesi ve görüntü, ses ve videoların sanal ortama aktarılmasını kolaylaştıran uygulamaların gelişmesiyle birlikte insanların kişisel bilgi, video ve görsellerine ulaşmak çok daha kolay hale gelmiştir. Derin sahte teknolojisi, derin öğrenme ve yapay zekâ tekniklerini kullanarak gerçek görüntü veya seslerin sahtelerini üretmek için kullanılmaktadır. Günümüzde eğlence ve film endüstrilerinde kullanılmasının yanı sıra, sahte haber oluşturma ve insanları itibarsızlaştırma gibi durumlarda da kullanılmaktadır. Bu durumların önüne geçmek için literatürde derin sahte görsel ve videoların tespitine yönelik farklı çalışmalar yapılmıştır. Bu çalışmada kapsamlı bir literatür taraması yapılmış ve farklı veri setlerinden veya videolardan gerçek ve sahte görseller toplanmış, etiketlenmiş ve gerekli ön işleme adımları uygulanarak bir veri seti oluşturulmuştur. Oluşturulan veri seti ile Evrişimli Sinir Ağlarını kullanarak geleneksel yöntemlerden farklı bir şekilde sınıf olasılıklarını hesaplayan ve tüm işlemleri tek bir regresyon probleminde ele alan hızlı ve yüksek doğrulukla tespit yapabilen YOLOv3 teknolojisi ile eğitim gerçekleştirilmiş ve modelleme süreci anlatılmıştır. Çalışmada yapılan testlerle derin sahte teknolojisiyle üretilmiş sahte görüntüleri 95 doğrulukla tespit edebilen bir model elde edilmiştir.

Makale detayları

A YOLOV3-Based Method for Detecting Deepfake Manipulated Facial Images

Yazarlar:

Mehmet Karaköse | Hasan Yetiş | Mert Çeçen

Yayıncı:

Fırat University

Ana Başlıklar:

Deep Learning

Dergi:

Turkish Journal of Science and Technology Vol.19 Issue 2, pp.315-324

Belge Tipi:

Makale

Yayın Tarihi:

9/2024

Belge Dili:

İngilizce

Document ID (Org No.):

2001328531

pISSN:

1308-9080

Convolutional neural networks

deepfake image detection

deep learning

YOLOV3Evrişimli sinir ağları

deepfake görüntü algılama

derin öğrenme

YOLOV3

Sayı

Makale detayları

Yazarlar: Mehmet Karaköse | Hasan Yetiş | Mert Çeçen

Yayıncı: Fırat University

Ana Başlıklar: Deep Learning

Dergi: Turkish Journal of Science and Technology Vol.19 Issue 2, pp.315-324

Belge Tipi: Makale

Yayın Tarihi: 9/2024

Belge Dili: İngilizce

Document ID (Org No.): 2001328531

pISSN: 1308-9080

Sayı

Yazarlar:

Mehmet Karaköse | Hasan Yetiş | Mert Çeçen

Yayıncı:

Fırat University

Ana Başlıklar:

Deep Learning

Dergi:

Turkish Journal of Science and Technology Vol.19 Issue 2, pp.315-324

Belge Tipi:

Makale

Yayın Tarihi:

9/2024

Belge Dili:

İngilizce

Document ID (Org No.):

2001328531

pISSN:

1308-9080