This research focuses specifically on GAN-generated synthetic faces, which underpin many face-centric deepfakes, and investigates efficient detection approaches using image analysis alone, suggesting that combining hierarchical CNN features with shifted-window self-attention provides an efficient and computationally lightweight method for detecting GAN-generated synthetic faces.
Abstract
Modern generative models, such as GANs, diffusion architectures, and autoregressive systems, now produce facial images that are nearly indistinguishable from authentic photographs. This capability makes detecting forged images increasingly difficult, raising serious concerns about identity theft, fraud, and misinformation campaigns. Our research focuses specifically on GAN-generated synthetic faces, which underpin many face-centric deepfakes, and investigates efficient detection approaches using image analysis alone. Existing detection systems rely heavily on either convolutional neural networks (CNNs) or global vision transformers. While CNNs excel at identifying texture-based local features, they struggle with broader contextual understanding. Traditional Vision Transformer (ViT) models can capture long-range structures effectively, but demand substantial computational resources. Our work explores Swin-Transformer-based architectures across three implementations: a compact Swin Transformer trained from the ground up, ImageNet-1K pre-trained Swin-Tiny and Swin-Small models adapted for binary classification, and a novel hybrid combining EfficientNet-B0's convolutional processing with a Swin Transformer backend. We evaluated all models using the 140K Real and Fake Faces dataset, which includes StyleGAN-generated fake faces alongside authentic images from Flickr and DFDC, with balanced splits for training, validation, and testing. The EfficientNetB0+Swin hybrid achieved 99% accuracy and a 99.44% recall on 5,000 test images, outperforming both pure Swin variants and a previous CNN-only baseline on this dataset. Our results suggest that combining hierarchical CNN features with shifted-window self-attention provides an efficient and computationally lightweight method for detecting GAN-generated synthetic faces.
Modern generative models, such as GANs and diffusion models, produce highly realistic synthetic images that are virtually indistinguishable from authentic photographs, posing significant risks in forensics, journalism, and data verification. To address the need for a fast and accessible detection tool, this paper propo...
K. Dukhnovska, T. I. Holub· Information Technologies and...· 0 citations
Recent advances in Generative Adversarial Network (GAN)-based image synthesis and autoencoder-driven facial manipulation have raised serious concerns about the trustworthiness of visual media. This paper presents a compact yet effective deepfake detection framework built on a Convolutional Neural Network (CNN) trained...
K. S. R. Krishna, M. S. Kasyapa, D. Dinesh et al.· International Conference on...· 0 citations
This work introduces a novel architecture by employing RGB and ELA of an image as a parallel stream of input for the two encoders (a lightweight neural network), primarily designed for localizing image manipulations that introduce compression inconsistencies, particularly image splicing and retouching.
A. Jaiswal, Asad Nizami, R. Srivastava· Neural computing & applicati...· 0 citations
The rapid advancement of deep generative models, especially Generative Adversarial Networks (GANs) and Diffusion Models, has escalated the creation of highly realistic synthetic media, posing significant threats to information security through misinformation and fraud. The core of current detection methodologies encomp...
Tian-Hua Tang· ITM Web of Conferences· 0 citations
Recent progress in generative deep learning—progressing from early autoencoders and GANs to modern latent
diffusion models and neural vocoders—have made synthetic media generation accessible, realistic, and increasingly difficult
to identify. While these tools offer productive uses in creative production and accessib...
Chayadevi M. L., Harsha Bharadwaj, Guru Raj Patil et al.· International Journal of Inn...· 0 citations
This paper introduces HybridLivenessNet, a novel deep neural architecture that combines MobileNetV3-Small as a computationally efficient convolutional feature extractor with a multi-head Transformer encoder, thus capturing diverse and complementary representations at the local texture and global spatial levels.
Shivakumar Dalali· Natural Resources for Human...· 0 citations
Known for his clear and elegant writing style, Bertsekas shaped fields from control and optimization to large-scale computation and artificial intelligence.