Topological Data Analysis for Image Tampering Detection

Abstract

This paper introduces a topological approach to detection of image tampering for forensics purposes. This is based on the emerging Topological Data Analysis (TDA) concept of persistent homological invariants associated with certain image features. Image features of interest are pixels that have a uniform Local Binary pattern (LBP) code representing texture feature descriptors. We construct the sequence of simplicial complexes for increasing sequence of distance thresholds whose vertices are the selected set of pixels, and calculate the corresponding non-increasing sequence of homology invariants (number of connected components). The persistent homology of this construction describes the speed with which the sequence terminates, and our tamper detection scheme exploit its sensitivity to image tampering/degradation. We test the performance of this approach on a sufficiently large image dataset from a benchmark dataset of passport photos, and show that the persistent homology sequence defines a discriminating criterion for the morphing attacks (i.e. distinguishing morphed images from genuine ones).

Face Recognition Systems Under Morphing Attacks: A Survey

“…The features extracted by texture descriptors can be further processed. A more complex method for morphing detection is proposed in [75] and [76], where a Vietoris-Rips complex is built of the responses of uniform LBP extractors on the image. In [100], a high detection performance was shown for a linear SVM trained on high-dimensional LBP features [105] extracted from the FEI database [87].…”

Detection of Face Morphing Attacks Based on PRNU Analysis

“…A more complex method for morphing attack detection is proposed in [16], [17], where a Vietoris-Rips complex is formed from the reactions of uniform LBP extractors on the image. In [38] a high detection performance was shown by 1 Also the morphed images used in this work can not be published due to licensing conditions as these are generated based on subsets of available image database collected by different institutions.…”

Face morph detection for unknown morphing algorithms and image sources: a multi‐scale block local binary pattern fusion approach

“…An extension of these algorithms to several colour channels [13, 14] or higher dimensions [22] can lead to further improvements. Other texture descriptors, such as unified LBPs (ULBPs) [17, 18] or weighted local magnitude patterns (WLMPs) [16] have also been tested.…”