Blind Approach for Digital Image Forgery Detection

Abstract

In the digital era of where everyone is exposed to a visual imagery in very large extent. Digital images are very convincible way to share information. Due to the rapidly growing field of digital image acquirement and editing software that are impressive as well sophisticated with many advanced features. Manipulation with features of digital image can perform easily with the help of editing tools, which are cost effectively available online or offline and do not leave any visible footprint of tampering with an image. Forgery with the digital image is an unavoidable problem concern with the image authenticity and also with image integrity. Which raising a compulsion to take an immediate action on the forgery of the digital image to verify the authenticity and maintain the integrity. To encounter the problem of authenticity of digital image, this paper proposed a methodology for detection of image splicing forgery using the blind approach i.e., passive method to detect the spliced region in the digital image. In passive approach, there is no provision for the pre-introduction of the watermark and preembedded digital signature during the time of image obtainment. This paper mainly concern with the image splicing forgery and it initiate with the DWT (Discrete Wavelet Transform) method, which will decompose the image into sub images and obtain coefficient for each sub image. After that for feature extraction we will use SURF (Speed-Up Robust Features) and finally SVM (Support Vector Machine) will perform classification for splicing forgery detection in digital image.

A Deep Learning Model to Inspect Image Forgery on SURF Keypoints of SLIC Segmented Regions

“… Image morphing: forming a new image that is completely different from two previous images. The necessary effects are made to harmonize the two images with each other [12].…”

Copy Move Forgery Detection: The Current Implications and Contemporary Practices

“…Thus the resulting great majority of image forgery detection algorithms employ a passive-based method to detect the tampering discussed in this study. There are two forms of passive detection technology: block based methods and key point based strategies [8]. The novel study as presented in this paper is divided into multiple sensations which further elaborates on the conventional approaches, the current standard practices and cutting edge methodologies, and finally the conclusion and the possible future work and related implications.…”

2018 Media Forensics Challenges (MFC18) :

“…However, previous studies have been limited to a single manipulation evaluation to measure the accuracy and robustness of the systems in the forensic applications. In spite of the wide variety of possible manipulation operations on the tampered media, the evaluations were often focused on a single processing manipulation [5,7,8]. Most widespread manipulation methods were adding (e.g., splicing), replicating (e.g., cloning), or removing (e.g., seam carving) content.…”