Colorization image detection method based on color statistical differences

Abstract

The invention provides a colorization image detection method based on color statistical differences. The method comprises steps of firstly, constructing the statistical distribution map of the training image set and the statistical distribution map of all the training images in the training image set; calculating the detection characteristic parameter based on the statistical distribution map of the training image set according to the difference maximization principle, calculating the statistical distribution map of each training image based on the detection characteristic parameter, and calculating the distribution characteristic of the statistical distribution map of each training image to construct the effective complete detection characteristics of the training images; and finally, using the complete detection characteristics of the training images and the corresponding training image tags to train the classifier to obtain an effective detection classifier. Therefore, the complete detection characteristic can be extracted for the image to be detected, so that the trained detection classifier can be used for detection.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a dyeing and forgery image detection method based on color statistical difference. Background technique [0002] With the rapid development of multimedia technology, the birth of a large number of image modification tools makes it easier to modify / forge images. Therefore, many image forgery incidents have occurred in recent years. According to the collection of international image counterfeiting expert HaniFarid and his team, forged images exist in many news events such as political scandals and weapons and equipment propaganda. At the same time, forged images may also exist in case evidence and even a large number of personal photos on the Internet. It can be seen that forged images have had a serious negative impact on the credibility of news, the handling of cases by public prosecutors, and even the trust between people. [0003] At present, there are mai...

AI Technical Summary

Problems solved by technology

However, there is currently no detection technology for dyed and forged images. It is too time-consuming and labor-intensive to use a special person to detect with the naked eye, and it is not suitable for a large amount of data. Traditional passive detection methods usually rely on different characteristics. Assumptions make it impossible to detect these generative forged images universally and efficiently. Therefore, research on dyeing and forgery image detection methods can fill in the gaps in the field of forged images, and effectively help users detect suspected dyeing forgery images.

Images