Image splicing tampering positioning method based on full convolutional neural network

Abstract

The invention discloses an image splicing tampering positioning method based on a full convolutional neural network. The method includes: establishing a splicing tampered image library; initializing an image splicing tampering positioning network based on a full convolutional neural network, and setting a training process of the network; initializing network parameters; reading a training image, performing training operation on the training image, and outputting a splicing positioning prediction result of the training image; calculating an error value between a training image splicing positioning prediction result and a real label, and adjusting network parameters until the error value meets a precision requirement; performing post-processing on the prediction result of which the precisionmeets the requirement by using a conditional random field, adjusting network parameters, and outputting a final prediction result of the training image; and reading the test image, predicting the test image by adopting the trained network, carrying out post-processing on a prediction result through a conditional random field, and outputting a final prediction result of the test image. The methodhas high splicing tampering positioning precision, the network training difficulty is low, and the network model is easy to converge.

Description

technical field [0001] The invention belongs to the technical field of deep learning, and in particular relates to an image splicing and tampering positioning method. Background technique [0002] With the development of science and technology, digital images are widely used in news, business, medical imaging, judicial criminal investigation and other fields. On the other hand, various image processing software represented by Photoshop, ACDSee, and Freehand can be used proficiently by more and more non-professionals, and the authenticity of digital images is facing severe challenges. Passive forensics in digital image forensics has more practical value and practical significance because it does not require any preprocessing. Among the various methods of tampering, splicing is the most commonly used method. Specifically, it refers to the operation of copying a part of an image from one image and then pasting it into another image, which will have a greater impact on the cont...

AI Technical Summary

Problems solved by technology

Liu et al. proposed to use fully convolutional neural network (FCN) instead of CNN, and combined with conditional random field (CRF) to fuse the positioning results of three FCNs. Although the accuracy has been improved, CRF and FCN are independent and are not an end-to-end structure.

Chen et al. changed the use of CRF, enhanced the learning of the target area, and made the entire network an end-to-end learning system. The positioning accuracy was improved, but as the depth of the network model deepened, the training speed of the model continued to decrease. It is becoming more and more difficult to train the ideal model in the data set

[0005] Based on the above, the existing traditional image stitching and tampering positioning methods have the problem that the positioning accuracy needs to be improved, and the positioning method based on deep learning has the problem of difficulty in training deep and complex models.

Images