A Double Compressed Video Tampering Detection Method Based on Blocking Effect

Abstract

The invention discloses a method for tampering detection of double-compressed video based on block effect, which uses block effect as a feature to perform tamper detection, including the following steps: Step 1: perform grayscale processing on the video to be tested and extract a feature of block effect; Step 2: Select different secondary compression parameters for the video to be tested to compress to obtain multiple compressed videos; Step 3: Extract the blockiness features of the compressed video, and construct a feature vector; Step 4: Calculate and fit the feature vector ; Step 5: compare the obtained fitting result with the fitting threshold for detection. The detection method provided by the invention adopts block effect features, can detect tampered videos with more diverse compression methods, and provides an effective new method for double compressed video detection.

Description

technical field [0001] The invention relates to a double-compression video tampering detection method, in particular to a block effect-based double-compression video tampering detection method. Background technique [0002] In the new multimedia era, simple video editing software on the Internet is almost overflowing, which causes the integrity and security of existing video content to be threatened. Today's video is used as a kind of judicial evidence. If criminals master tampering technology and tamper with evidence video to achieve illegal purposes such as evading judicial disposal or blaming others, it will have a bad impact on fair justice and social security. Therefore, passive detection of video tampering emerges at the historic moment, and set off a research boom at home and abroad. Double compressed video detection is a kind of passive detection, which provides a new idea for detecting video tampering caused by frame deletion and frame insertion operations. [000...

AI Technical Summary

Problems solved by technology

Based on the detection of DCT coefficients, according to the DCT coefficient matrix obtained after quantization, there are certain rules in it, such as the probability distribution of the first digit of the AC coefficient, and the DCT coefficient histogram presents convex function characteristics, etc. to detect, but for two compression There are certain constraints on the quantization parameter setting of

The detection based on the GOP structure is detected from the change of some characteristics of the video caused by the change of the GOP structure, such as detection by detecting that there is a certain periodicity in the motion error estimation after the tampering, etc. This method is easy to understand, but lacks a final identification The standard of test results often needs to be judged by naked eyes

Images