New dissimilarity measures for image phylogeny reconstruction

Costa, Filipe; Oliveira, Alberto; Ferrara, Pasquale; Dias, Zanoni; Goldenstein, Siome; Rocha, Anderson

doi:10.1007/s10044-017-0616-9

Cited by 9 publications

(6 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Techniques for associating images in a general way include comparing global image representations [23], [24], employing bags of visual features [25], [26] and using convolutional neural networks (CNNs) [27]- [30]. Techniques for associating images in a specialized way include assessing local feature matching [31]- [35], image patch matching [36], and evaluating the quality of image registration, color matching, and mutual information [37], [38]. Particularly, provenance analysis is by definition closer to the specialized tasks; for that reason, in this work, we benefit more from techniques mentioned in the latter group.…”

Section: Related Workmentioning

confidence: 99%

“…Alternatively, Oikawa et al [46] propose the use of clustering techniques for finding the various phylogeny trees; the idea is to group images coming from the same source, while placing semantically similar images in different clusters. Finally, Costa et al [37] improve the creation of the dissimilarity matrices, regardless of the graph algorithm used for constructing the trees.…”

Section: Related Workmentioning

confidence: 99%

“…The p(x, y) value regards the joint probability distribution function of R 1 and R 2 . As explained in [37], it can be approximated by: p(x, y) = h(x, y)…”

Section: Mi-based Dissimilaritymentioning

confidence: 99%

See 2 more Smart Citations

Image Provenance Analysis at Scale

Moreira¹,

Bharati²,

Brogan³

et al. 2018

Preprint

Self Cite

View full text Add to dashboard Cite

Prior art has shown it is possible to estimate, through image processing and computer vision techniques, the types and parameters of transformations that have been applied to the content of individual images to obtain new images. Given a large corpus of images and a query image, an interesting further step is to retrieve the set of original images whose content is present in the query image, as well as the detailed sequences of transformations that yield the query image given the original images. This is a problem that recently has received the name of image provenance analysis. In these times of public media manipulation (e.g., fake news and meme sharing), obtaining the history of image transformations is relevant for fact checking and authorship verification, among many other applications. This article presents an end-to-end processing pipeline for image provenance analysis, which works at real-world scale. It employs a cutting-edge image filtering solution that is custom-tailored for the problem at hand, as well as novel techniques for obtaining the provenance graph that expresses how the images, as nodes, are ancestrally connected. A comprehensive set of experiments for each stage of the pipeline is provided, comparing the proposed solution with state-of-the-art results, employing previously published datasets. In addition, this work introduces a new dataset of real-world provenance cases from the social media site Reddit, along with baseline results.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Image Provenance Analysis at Scale

Moreira¹,

Bharati²,

Brogan³

et al. 2018

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Thus, traditionally, the first step of asymmetric measure computation involves geometric registration, followed by color channel normalization and compression matching [10]. Other works focus on using waveletbased denoising technique [19], and a combination of gradient estimation and mutual information techniques [20] to derive an improved asymmetric measure.…”

Section: Related Workmentioning

confidence: 99%

Face Phylogeny Tree Using Basis Functions

Banerjee

Ross

2020

IEEE Trans. Biom. Behav. Identity Sci.

View full text Add to dashboard Cite

Photometric transformations, such as brightness and contrast adjustment, can be applied to a face image repeatedly creating a set of near-duplicate images. Identifying the original image from a set of such near-duplicates and deducing the relationship between them are important in the context of digital image forensics. This is commonly done by generating an image phylogeny treea hierarchical structure depicting the relationship between a set of near-duplicate images. In this work, we utilize three different families of basis functions to model pairwise relationships between near-duplicate images. The basis functions used in this work are orthogonal polynomials, wavelet basis functions and radial basis functions. We perform extensive experiments to assess the performance of the proposed method across three different modalities, namely, face, fingerprint and iris images; across different image phylogeny tree configurations; and across different types of photometric transformations. We also utilize the same basis functions to model geometric transformations and deep-learning based transformations. We also perform extensive analysis of each basis function with respect to its ability to model arbitrary transformations and to distinguish between the original and the transformed images. Finally, we utilize the concept of approximate von Neumann graph entropy to explain the success and failure cases of the proposed IPT generation algorithm. Experiments indicate that the proposed algorithm generalizes well across different scenarios thereby suggesting the merits of using basis functions to model the relationship between photometrically and geometrically modified images. !• Both the authors are with the

show abstract

“…The procedure is similar to the dissimilarity estimation detailed in Section 2.1.1, with two main differences: no alignment is necessary (as it was already done in the previous step) and only the pixels of R are used to estimate the color transformations. The color transformations are estimated using histogram matching, as proposed by Costa et al [16]. Compression is once again estimated by compressing the transformed I A using I B 's quantization table.…”

Section: Local Dissimilarity Estimation;mentioning

confidence: 99%

Multiple parenting relations in image phylogeny

Oliveira¹

View full text Add to dashboard Cite

Due to the large amount of images shared on the web, tracking the spread and evolution of their content became an increasingly important problem. Recently, several works have been concerned with modeling the parenthood relationships between near duplicates in a set of images. Two images share a parenthood relationship if one is obtained by applying transformations to the other. However, this is not the only form of parenting that can exist among images. An image might be a composition created through the combination of the semantic information existent in two or more source images, establishing a relationship between the sources and the composite. We name as Multiple Parenting Phylogeny the problem of identifying such relationships in a set containing near-duplicate subsets of source and composition images. To tackle this problem, we present in this dissertation a three-step solution: (1) separation of near-duplicate groups; (2) classification of the relations between the groups; and (3) identification of the images used to create the original composition. Moreover, next we extend upon this framework by introducing key improvements, such as better identification of when two images share content, and improved ways to compare this content. We evaluate the proposed approaches in a scenario containing three subsets of near duplicates: one belonging to the composition image, and two to the source images used to create it. Three datasets were developed for this work, two with automatically created compositions and one of realistic, professionally-made compositions. The original and extended frameworks are compared by means of quantitative metrics established for evaluating the accuracy in reconstructing phylogenies and finding multiple parenting relationships. Finally, we discuss some particularities of the framework, such as the importance of an effective reconstruction of phylogenies and how more complex compositions can negatively affect the precision of the method.

show abstract

New dissimilarity measures for image phylogeny reconstruction

Cited by 9 publications

References 40 publications

Image Provenance Analysis at Scale

Image Provenance Analysis at Scale

Face Phylogeny Tree Using Basis Functions

Multiple parenting relations in image phylogeny

Contact Info

Product

Resources

About