Enhancing Image Forgery Detection Using 2-D Cross Products
The availability of sophisticated, easy-to-use image editing tools means that the authenticity of digital images can no longer be guaranteed. This chapter proposes a new method for enhancing image forgery detection by combining two detection techniques us
- PDF / 537,809 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 87 Downloads / 221 Views
The availability of sophisticated, easy-to-use image editing tools means that the authenticity of digital images can no longer be guaranteed. This chapter proposes a new method for enhancing image forgery detection by combining two detection techniques using a 2-dimensional cross product. Compared with traditional approaches, the method yields better detection results in which the tampered regions are clearly identified. Another advantage is that the method can be applied to enhance a variety of detection algorithms. The method was tested on the CASIA TIDE v2.0 public dataset of color images and the results compared against those obtained using the re-interpolation, JPEG noise quantization and noise estimation techniques. The experimental results indicate that the proposed method is efficient and has superior detection characteristics.
Keywords: Image tampering, forgery detection, cross product
1.
Introduction
The proliferation of low-cost, high-quality digital cameras and sophisticated image processing software make it very easy to manipulate or forge digital images without any obvious traces. Due to the dramatic increase in doctored images [1], the authenticity and trustworthiness of digital images are always in question. This situation can pose serious problems in criminal investigations, judicial proceedings, journalism, medical imaging and even insurance claim processing, where the authenticity of every digital image must be guaranteed. A digital image may be tampered with via image retouching, splicing and/or copy-move forging. Retouching, cloning and healing are methc IFIP International Federation for Information Processing 2016 Published by Springer International Publishing AG 2016. All Rights Reserved G. Peterson and S. Shenoi (Eds.): Advances in Digital Forensics XII, IFIP AICT 484, pp. 297–310, 2016. DOI: 10.1007/978-3-319-46279-0 15
298
ADVANCES IN DIGITAL FORENSICS XII
Figure 1.
Original and forged images [1, 3].
ods of image manipulation in which some elements are removed, altered, blurred or emphasized using parts or properties of the same image; this type of manipulation also involves the adjustment of some image properties (e.g., color, white balance and contrast). Splicing [14] is a common image tampering technique; the technique combines image fragments from the same or different images to create a new image. Another popular technique for manipulating images is copy-move forgery [2]; this technique duplicates certain parts of a target image and places them elsewhere in the same image, the objective being to hide or emphasize parts of the target image. Figure 1 shows an original image (left) and its forged counterpart (right) [1, 3]. A number of researchers have studied the problem of image forgery detection. Zhao et al. [15] have leveraged JPEG compression characteristics to detect image inpainting in JPEG images. Kaur and Jyoti [7] have developed an image tampering detection method based on the inconsistency of JPEG grids in a suspect image. Cao et al. [2] have proposed an algorithm for
Data Loading...
