.png){kind=icon}
.png){kind=icon}
Return
Differentiating Between Images Using Wavelet-Based Transforms: A Comparative Study
.png){kind=icon}
PRE
AI
DOI：10.1109/TIFS.2011.2162830.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We propose statistical image models for wavelet-based transforms, investigate their use, and compare their relative merits within the context of digital image forensics. We consider the problems of 1) differentiating computer graphics images from photographic images, 2) source camera and source scanner identification, and 3) source artist identification from digital painting samples. The features obtained from ridgelet and contourlet transform-based image models almost always perform better than the features obtained from wavelet-based image models for the problems at hand. We outline properties of efficient image representation, relate these properties to wavelet-based transforms, and discuss the experimental results in relation to the model properties.
Keywords:
Classification
contourlet transform
digital image forensics
feature selection
ridgelet transform (RIT)
source identification
support vector machines (SVMs)
wavelet transform
Journal
IF:
8
Papers:
5.2K
Citations:
2.3W