Image tampering detection method, terminal equipment and storage medium

Abstract

The invention relates to an image tampering detection method, terminal equipment and a storage medium, and the method comprises the steps: S1, collecting an untampered image and a corresponding tampered image, subtracting the tampered image from the untampered image to obtain a difference image of the tampered image, and enabling the untampered image and the tampered image to jointly form a training set; S2, constructing a binary classification network model, and training the binary classification network model through the training set, so that the trained binary classification network model can distinguish whether the image is tampered or not, wherein the binary classification network model comprises a feature extraction layer, an image attention layer and a classifier; S3, recognizing whether the image is tampered or not through the trained binary classification network model. The difference image is used as a real label. A real label is used as tampering supervision information, a neural network is used for training a binary classification network, the network is guided to obtain an accurate tampering detection probability value and a fake positioning graph, and the performanceof face tampering image recognition and classification can be effectively improved.

Description

technical field [0001] The invention relates to the field of image monitoring, in particular to an image tampering detection method, a terminal device and a storage medium. Background technique [0002] With advances in image editing technology and user-friendly editing software, low-cost tampering with the image generation process has become ubiquitous. But these advances have also lured real fake content that has a major impact on people's lives. They can be used to generate fake faces, spread on ubiquitous social networks, and even be used to crack face recognition systems, causing serious harm to security. Therefore, it is crucial to develop methods that can reliably detect such manipulations, and in recent years there has been increasing attention to this goal. [0003] Face tampering detection methods are mainly divided into two types: traditional methods and deep learning-based methods. The traditional method is to design manual features according to a certain type...

AI Technical Summary

Problems solved by technology

[0004] Although the current tamper detection methods based on traditional methods and deep learning methods can achieve certain detection results, there are still defects

Most of the traditional methods are designed for a certain tampering method, so the robustness and scalability are poor, and they are not suitable for various tampering detection

However, most of the methods based on deep learning are for tamper detection, with weak or no positioning ability, and the accuracy is limited to a certain extent.

Images