A detection method for jpeg image quantization table tampering

Abstract

The invention discloses a method for detecting tampering of quantization tables of JPEG pictures, which comprises the following steps: selecting several JPEG pictures as original JPEG pictures, tampering the quantization tables of the original JPEG pictures, and obtaining the JPEG pictures after the quantization tables have been tampered with; and quantization table tampered JPEG picture feature extraction; according to the feature extracted from the original JPEG picture and quantization table tampered JPEG picture, the classifier is trained to obtain a classifier model; extract the features of the picture to be tested, according to the classifier model Determine whether the quantization table of the image to be tested has been tampered with. The invention can be used as a technology for identifying JPEG pictures of tampered quantization tables, provides an effective automatic means for detecting malicious image tampering, and has broad application prospects in judicial and criminal investigation and evidence collection.

Description

technical field [0001] The invention belongs to the field of multimedia information security and evidence collection, and in particular relates to a method for detecting tampering of a JPEG picture quantization table. Background technique [0002] With the popularization of various powerful professional image editing software such as Photoshop and ACDsee, tampering with images has become easier and easier. Even ordinary users can tamper with visually enough fake pictures. However, after the original JPEG image has been tampered with by Photoshop, ACDsee, etc., it will inevitably introduce its own quantization table when it is saved as JPEG again, because the quantization table of the software is mostly different from the quantization table of the JPEG image taken by the camera and mobile phone, which leads to The quantization tables of JPEG pictures before and after tampering are inconsistent. If the quantization table is not replaced back to the quantization table before t...

AI Technical Summary

Problems solved by technology

[0005] In terms of passive image authentication, although the originality of a JPEG image can be judged theoretically by detecting whether the JPEG image has undergone two JPEG compressions, due to the quantization table of the JPEG image generated by the current camera, mobile phone and other shooting devices and the commonly used quantization of software such as Photoshop, etc. Tables are similar, so the detection results of two JPEG compressions are usually not ideal

Images