An image resampling operation detection method

Abstract

The invention relates to an image resampling operation detection method. The method comprises the steps of representing image resampling operation by using a model; carrying out initial blind deconvolution operation on a resampled image to obtain an initial kernel; working out initial kernels with different sizes so as to obtain a group of convolution kernel sets with multiple scales and different sizes; comparing a result, obtained by carrying out convolution on one convolution kernel in the convolution kernel sets and an original image corresponding to the convolution kernel, with the resampled image, wherein the convolution kernel with the minimum mass difference is an optimal kernel, namely, system output; unfolding the optimal kernel to obtain a column vector; putting the obtained column vector into in a classifier for training a model to obtain a classification model; repeating the step 2 to the step 5 for the image to be detected, and putting the obtained column vector into the classification model obtained in the step 6 for testing so as to obtain a final detection result. The image resampling operation detection method can effectively reduce the training time of the support vector machine (SVM) classifier. Furthermore, the method in high in compression resistance, and has excellent detection accuracy rate.

Description

technical field [0001] The invention relates to an image tampering detection method. In particular, it relates to a method for detecting image resampling operations. Background technique [0002] With the advent of low-cost and high-resolution digital cameras and sophisticated editing software, digital images can be easily manipulated and changed. Fake digital images are often difficult to distinguish from real photos. Therefore, photographs can no longer serve as a record of the authenticity of events. Especially in the fields of judicial system and news media, the authenticity of image information plays an increasingly important role. In order to ensure the authenticity and integrity of digital information, digital image passive forensics technology came into being. Different from the active forensics technology represented by digital signature and digital watermark, the passive forensics technology is only based on the acquired digital image to find out whether there ...

AI Technical Summary

Problems solved by technology

[0005] Although the detection of resampling has been carried out for many years, there are still many problems, such as poor resistance to JPEG compression, difficulty in detecting specific zoom ratios, and difficulty in detecting aperiodic interpolation, etc.

Images