Exposing Image Manipulation with Curved Surface Reflection

Objectives: One of the principal problem in image forensics is identifying if a particular image is manipulated or not. Nowadays powerful image editing software increases the difficulty in finding the authentic image. Our approach is to detect curved surface manipulated area inserted in the photograph by describing a geometrical technique. Methods/ Statistical Analysis: This technique is combined with multiple image segmentation methodologies. The main objective of our work is to identify the manipulated area from the reflected images. Findings: Our contribution of this paper is to design a new geometrical relationship between the original and fake images. We use the new notations for this geometrical representation and combine these notations to the basic image segmentation methods. Finally we identify the fake reflected image using the different approaches and compare the results and produce the best performance of our work. Application/Improvements: The experimental result shows that our method achieves promising performance for tampering the curved surface reflections. The result of our work, we compared the 100 images and produce the best PSNR value and compression ratio.