Font Size: a A A

Video Watermarking Algorithm AVS DCT Domain

Posted on:2014-01-30Degree:MasterType:Thesis
Country:ChinaCandidate:J L DingFull Text:PDF
GTID:2268330425951027Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
With the development of multimedia technology and computer network, the application ofdigital video becomes more and more popular. Consequently the infringement and piracy of thedigital video are becoming more and more serious. To solve the issue of copyright, videowatermarking is an important way with its effective performance and simple operation. And moreand more scholars pay their attentions on it. AVS video codec standard is the second generationsource coding standard which our country independently research through a lot of outstandingachievement of other scientific research institutions. It has the advantages of excellentperformance and simple patent authorization. This paper has researched on the AVS videowatermark algorithm based on DCT domain, mainly includes the following aspects:⑴In order to enhance the ability against brute force attack for the image, this paper usesoffset variable m and parameter matrix as the key to improve the performance of Arnoldtransform. Compared with the traditional algorithm which uses scrambling times as the key, thealgorithm which the paper presented has increased the key space of image scrambling effectively.⑵This paper choose I frame which coding independently in AVS standard as the embeddedframe, the I frame macro block was partitioned to88block, according to the rules ofembedding, the watermark information embedded in low frequency part or high frequency part ofDCT quantified coefficients reasonably. Watermark extraction can realize blind extractionwithout the original video.⑶In the process of embedding watermark information in high frequency, This paperintroduces regulatory factor p to adjust the amplitude of the local exchange energyE ACin88block. It can realize automatic adjustment for embedding progress. This paper has modified of thequantized coefficients according to the texture information, the edge information and the activityof the image. It combined the embedding process with the human visual characteristic to enhancethe adaptability of the watermark.⑷The algorithm this article proposed has been simulated for lots of brute force attacks,including anti repeated quantization, frame shear, frame deletion, noise attacks and so on. Theexperimental results demonstrated that the algorithm has good robustness and invisibility.
Keywords/Search Tags:AVS, Video frequency watermark, DCT coefficient, Blind examination
PDF Full Text Request
Related items