Deep forgery discriminator via image degradation analysis

Abstract

Generative adversarial network-based deep generative model is widely applied in creating hyper-realistic face-swapping images and videos. However, its malicious use has posed a great threat to online contents, thus making detecting the authenticity of images and videos a tricky task. Most of the existing detection methods are only suitable for one type of forgery and only work for low-quality tampered images, restricting their applications. This paper concerns the construction of a novel discriminator with better comprehensive capabilities. Through analysis of the visual characteristics of manipulated images from the perspective of image quality, it is revealed that the synthesized face does have different degrees of quality degradation compared to the source content. Therefore, several kinds of image quality-related handicraft features are extracted, including texture, sharpness, frequency domain features, and deep features, to unveil the inconsistent information and modification traces in the fake faces. In this way, a 1065-dimensional vector of each image is obtained through multi-feature fusion, and it is then fed into RF to train a targeted binary classification detector. Extensive experiments have shown that the proposed scheme is superior to the previous methods in recognition accuracy on multiple manipulation databases including the Celeb-DF database with better visual quality.

MSFRNet: Two‐stream deep forgery detector via multi‐scale feature extraction

“…Earlier works utilize various hand-crafted descriptors followed by a simple binary classifier to capture the spoofing patterns. For example, our previously work [16] extracts several manual features related to image quality, including textural features, sharpness measurements, frequency domain features, and deep features, to expose manipulation traces left by fake face, and finally a random forest (RF) model is employed for classification. Reshma et al [17] adopt textural and gradient features with the help of support vector machine (SVM) for face liveness detection.…”

The construction of urban cultural and creative industries using deep learning and information management

“…Through extensive exploration of deep learning and information management, CCI have leveraged these technologies to foster the emergence of diverse downstream sectors. For instance, deep forgery technology employs deep learning-based face generation and manipulation techniques to synthesize authentic-looking forged faces, thereby catalyzing advancements in the entertainment and cultural industry [ 24 ]. Similarly, the utilization of CNN-enabled shot boundary detection and boundary frame target detection technology in the recording of intangible cultural heritage videos facilitates the propagation, preservation, and safeguarding of intangible cultural heritage [ 25 ].…”

A novel blind tamper detection and localization scheme for multiple faces in digital images

“…Over the years, an enormous number of personal images have been shared and distributed through the internet and social media networks which are exposed to the danger of unauthorized manipulations and attacks. The manipulated images can be used for unethical aims such as financial fraud, causing political conflicts by spreading fake news, damaging the reputation of successful companies, celebrities, political leaders, and many others [3,4]. Generally, the image forgery detection field witnessed a lot of interest because of its importance in revealing manipulations in digital images [5].…”