作者: Dajun He , Qibin Sun , Qi Tian
DOI: 10.1155/S1110865704312217
关键词: Artificial intelligence 、 Theoretical computer science 、 Computer vision 、 Digital watermarking 、 Cryptographic hash function 、 Digital signature 、 Watermark 、 Segmentation 、 Video processing 、 Computer science
摘要: An object-based video authentication system, which combines watermarking, error correction coding (ECC), and digital signature techniques, is presented for protecting the authenticity between objects their associated backgrounds. In this a set of angular radial transformation (ART) coefficients selected as feature to represent object background, respectively. ECC cryptographic hashing are applied those generate robust watermark. This content-based, semifragile watermark then embedded into frame by before MPEG4 coding. embedding extraction, groups discrete Fourier transform (DFT) randomly selected, energy relationships employed hide extract The experimental results demonstrate that our system compression, segmentation errors, some common processing such translation, rotation, scaling while securely preventing malicious modifications. proposed solution can be further incorporated public key infrastructure (PKI).