Image forgery detection using structural similarity index (SSIM), oriented fast and rotation brief (ORB) and scale – invariant feature transform (SIFT)

Chandran, Thiasan (2021) Image forgery detection using structural similarity index (SSIM), oriented fast and rotation brief (ORB) and scale – invariant feature transform (SIFT). Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Image forgery detection using structural similarity index (SSIM), oriented fast and rotation brief (ORB) and scale – invariant feature transform (SIFT).pdf - Submitted Version Download (543kB)
	Text (Full Text) Image forgery detection using structural similarity index (SSIM), oriented fast and rotation brief (ORB) and scale – invariant feature transform (SIFT).pdf - Submitted Version Restricted to Repository staff only Download (2MB)

Abstract

Digital images are widely used in everywhere in the world. It is very helpful to mankind in solving many problems especially in the communication field. Nowadays, almost every human being is owning a smart device which is able to capture photos and videos easily. Almost millions of photos are uploaded every day in media socials such as on Instagram and Facebook. Previously, photos have been used only for educational purposes and as moments memories. There is no problem until the images are used genuinely without any crime involvement. On the other hand, the images are playing important roles in the crime scenes. Images are being used as photographic evidences in crime cases. So, it is important to maintain the integrity of the images. Unfortunately, the advancement of technologies has introduced various photo editing software such Adobe Photoshop which is able to modify the images and alter the contents easily. People are getting more skills on manipulating the original images where it directly effecting the integrity of an image. So, it is difficult for digital forensic department to identify between the original images and manipulated images. In this project, a system with 3 image forgery detection algorithms have been used to identify the similarities and differences between two images. The algorithms are Structural Similarity Index (SSIM), Oriented FAST and Rotated BRIEF (ORB) and Scale-Invariant Feature Transform (SIFT). This system is developed using Python as the programming language to integrate with those algorithms to perform image forgery detection. As a result, the system is able to identify the similarities and differences between two images in term of similarities score values and running time. Finally, this project concluded that the best image forgery detection for rotated images is ORB and for copy move forgery images is SSIM. This thesis helps to find out and recommend the best algorithms among SSIM, ORB and SIFT in terms of similarities score value, running time and detection areas.

Actions (login required)

View Item