Image Forensics: Detecting duplication of scientific images with manipulation-invariant image similarity
This addresses the issue of image manipulation in scientific publications, which lacks scalable solutions, by offering an automated detection method, though it appears incremental as it builds on existing Siamese network approaches.
The paper tackles the problem of detecting manipulated duplicates of scientific images, such as through rotation or editing, by proposing a Siamese CNN that maps images into a space where duplicates are close and unique images are far apart, achieving a threshold-based separation with distances ≤1 for duplicates and >1 for unique images.
Manipulation and re-use of images in scientific publications is a concerning problem that currently lacks a scalable solution. Current tools for detecting image duplication are mostly manual or semi-automated, despite the availability of an overwhelming target dataset for a learning-based approach. This paper addresses the problem of determining if, given two images, one is a manipulated version of the other by means of copy, rotation, translation, scale, perspective transform, histogram adjustment, or partial erasing. We propose a data-driven solution based on a 3-branch Siamese Convolutional Neural Network. The ConvNet model is trained to map images into a 128-dimensional space, where the Euclidean distance between duplicate images is smaller than or equal to 1, and the distance between unique images is greater than 1. Our results suggest that such an approach has the potential to improve surveillance of the published and in-peer-review literature for image manipulation.