Identification of Natural Images and Computer Generated Graphics Based on Hybrid Features

Identification of Natural Images and Computer Generated Graphics Based on Hybrid Features

Fei Peng (Hunan University, China), Juan Liu (Hunan University, China) and Min Long (Changsha University of Science and Technology, China)
DOI: 10.4018/978-1-4666-4006-1.ch002
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Examining the identification of natural images (NI) and computer generated graphics (CG), a novel method is proposed based on hybrid features. Since the image acquisition pipelines are different, some differences exist in statistical, visual, and noise characteristics between natural images and computer generated graphics. Firstly, the mean, variance, kurtosis, skew-ness, and median of the histograms of grayscale image in the spatial and wavelet domain are selected as statistical features. Secondly, the fractal dimensions of grayscale image and wavelet sub-bands are extracted as visual features. Thirdly, considering the shortage of the photo response non-uniformity noise (PRNU) acquired from wavelet based de-noising filter, a pre-processing of Gaussian high pass filter is applied to the image before the extraction of PRNU, and the physical features are calculated from the enhanced PRNU. In the identification, a support vector machine (SVM) classifier is used in experiments and an average classification accuracy of 94.29% is achieved, where the classification accuracy for computer generated graphics is 97.3% and for natural images is 91.28%. Analysis and discussion show that the method is suitable for the identification of natural images and computer generated graphics and can achieve better identification accuracy than the existing methods with fewer dimensions of features.
Chapter Preview
Top

Natural image is the reflection of natural light and recorded by image acquisition devices. Firstly, the light passes through the lens and subsequently is converted to electrical signal by an image sensor. Generally, image sensor acquires only one color of the red, green, or blue components (RGB) in each pixel and the other two colors are obtained by CFA interpolation algorithms. Then other operations such as white balance, JPEG compression and other post-processing operations will be done. Finally, the image is completed (Lukas & Fridrich, 2005a). The whole image processing is shown in Figure 1.

Figure 1.

Acquisition process of a natural image

Complete Chapter List

Search this Book:
Reset