Video tampering detection method based on compressed sensing semi fragile watermark

Abstract

The embodiment of the invention discloses a video tampering detection method based on a compressed sensing semi fragile watermark, which comprises the steps of: extracting an MPEG-2 (Moving Picture Experts Group) compressed video, extracting image characteristics of a frame I through a compressed sensing technology on the basis of a coding mode of a frame I image and a characteristics of a DCT (Discrete Cosine Transformation) coefficient, generating and embedding two watermarks of a content authorization watermark and an integrity watermark; carrying out initial check on video integrity, inter-frame taper detection on the video and intra-content authorization on the video, and finally outputting a detection result and carrying out MPEG-2 decompression on the video. By implementing the invention, the image characteristics of the frame I are extracted by using the compressed sensing technology on the basis of the coding mode of the frame I image and the characteristic of the DCT coefficient, a semi-fragile content authorization watermark is generated, and then embedded in medium-high frequency coefficients of each macro block of the frame I image; and based on a motion vector characteristic of a frame P, Hash operation is carried out a sequence number of the frame P, the obtained key value is used as the integrity watermark to be embedded in a motion vector of the frame P, thus intra-frame taper detection is realized in the frame I, and inter-frame taper detection is realized in the frame P.

Description

technical field [0001] The invention relates to the field of communication technology, in particular to a video tampering detection method based on compressed sensing semi-fragile watermarks. Background technique [0002] With the emergence of powerful multimedia editing software, the modification of multimedia information such as audio and video has become simple and easy. Once falsified and falsified multimedia information is widely used in official media, scientific discoveries, insurance and court evidence, it will bring serious consequences. Tamper detection is mainly to solve the problem of integrity and originality identification of digital multimedia data. The tampering detection methods for digital images are relatively mature, but the tampering detection methods for digital videos are relatively few. [0003] Aiming at the characteristics of malicious tampering of video and utilizing the continuity of content between frames, this paper proposes to detect whether ...

AI Technical Summary

Problems solved by technology

At present, a reversible semi-fragile watermarking algorithm for distinguishing between MPEG-2 compression and malicious tampering is proposed. It embeds the watermark into the coefficients after the video frame is divided into 8×8 blocks and transformed by DCT. Preliminary positioning of internal tampering, but the positioning accuracy is not high and cannot handle inter-frame tampering

At present, a semi-fragile reversible video watermarking algorithm using hash function for MPEG-4 content authentication is also proposed. It embeds two watermarks in the I frame of MPEG-4 encoding of YUV video. This algorithm can verify the video content. And achieve tampering positioning, but for inter-frame tampering, the algorithm cannot accurately locate the added or deleted frames

Images