Distributions of the DCT coefficient for Watermark Detection

Author

Du, Yao Gang ; Ying, Chen ; Feng, Wang L. ; Zheng, Yuan

Author_Institution

Beijing Electron. Sci. & Technol. Inst., Beijing, China

Volume

13

fYear

2010

fDate

22-24 Oct. 2010

Abstract

In this paper, we derive the distribution of the DCT coefficients for a fixed block variance, and invoke the central limit theorem to show that this is approximately Gaussian. Then, by allowing the block variance to have a statistical distribution itself, we analyze when the DCT coefficients would follow a Laplacian distribution. Such knowledge would be useful, for instance, in quantizer design, information hiding capacity and Watermark Detection.

Keywords

data encapsulation; discrete cosine transforms; quantisation (signal); signal detection; statistical distributions; watermarking; DCT coefficient; Laplacian distribution; central limit theorem; fixed block variance; information hiding capacity; quantizer design; statistical distribution; watermark detection; Discrete cosine transforms; Gaussian distribution; Laplace equations; Probability density function; Random variables; Watermarking; DCT coefficient; Laplacian distribution; information hiding capacity;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Application and System Modeling (ICCASM), 2010 International Conference on

Conference_Location

Taiyuan

Print_ISBN

978-1-4244-7235-2

Electronic_ISBN

978-1-4244-7237-6

Type

conf

DOI

10.1109/ICCASM.2010.5622674

Filename

5622674