With the development of deep learning, image manipulation techniques have become increasingly advanced and difficult to detect, leading to serious trust crises in fields such as news reporting, judicial forensics, and social media. In recent years, deep learning-based image manipulation localization methods have achieved significant progress. Most current methods focus on extracting non-semantic information, which is reasonable, but since most manipulation operations modify the semantic content of images, both semantic and non-semantic information are indispensable. Since the introduction of Vision Transformer (ViT), it has dominated the field of computer vision, with its Transformer architecture demonstrating powerful capabilities in extracting image features. However, few ViT-based methods have been applied to image manipulation localization, and their performance remains unsatisfactory. Therefore, we combine traditional vision models with advanced ViT, integrating both semantic and non-semantic information to construct a new image manipulation localization framework called ViT-DGA. The traditional vision model is used to extract features from different frequency domains of the image, while a ViT-based intermediate layer supplements the original image features, effectively guiding the model to learn multi-perspective features. Additionally, we conducted extensive experiments on various datasets. The results demonstrate that ViT-DGA not only achieves superior localization accuracy compared to state-of-the-art image manipulation localization methods but also exhibits strong robustness in common post-processing scenarios.

Structural image features are exploited to construct perceptual image hashes in this work. The image is first preprocessed and divided into overlapped blocks. Correlation between each image block and a reference pattern is calculated. The intermediate hash is obtained from the correlation coefficients. These coefficients are finally mapped to the interval [0, 100], and scrambled to generate the hash sequence. A key component of the hashing method is a specially defined similarity metric to measure the “distance” between hashes. This similarity metric is sensitive to visually unacceptable alterations in small regions of the image, enabling the detection of small area tampering in the image. The hash is robust against content-preserving processing such as JPEG compression, moderate noise contamination, watermark embedding, re-scaling, brightness and contrast adjustment, and low-pass filtering. It has very low collision probability. Experiments are conducted to show performance of the proposed method.

Motivated by the image rescaling estimation method proposed by Gallagher (2nd Canadian Conf. Computer & Robot Vision, 2005: 65-72), we develop an image rotation angle estimator based on the relations between the rotation angle and the frequencies at which peaks due to interpolation occur in the spectrum of the image's edge map. We then use rescaling/rotation detection and parameter estimation to detect fake objects inserted into images. When a forged image contains areas from different sources, or from another part of the same image, rescaling and/or rotation are often involved. In these geometric operations, interpolation is a necessary step. By dividing the image into blocks, detecting traces of rescaling and rotation in each block, and estimating the parameters, we can effectively reveal the forged areas in an image that have been rescaled and/or rotated. If multiple geometrical operations are involved, different processing sequences, i.e., repeated zooming, repeated rotation, rotation-zooming, or zooming-rotation, may be determined from different behaviors of the peaks due to rescaling and rotation. This may also provide a useful clue to image authentication.