scholarly journals Two iterative methods for reverse image filtering

2021 ◽  
Author(s):  
Alexander G. Belyaev ◽  
Pierre-Alain Fayolle
Keyword(s):  
Internal Structure ◽  
Iterative Methods ◽  
Image Filtering ◽  
Black Box ◽  
Original Image ◽  
Edge Preserving ◽  
Image Filters

AbstractWe consider the problem of recovering an original image $${\varvec{x}}$$ x from its filtered version $${\varvec{y}}={\varvec{f}}({\varvec{x}})$$ y = f ( x ) , assuming that the internal structure of the filter $${\varvec{f}}(\cdot )$$ f ( · ) is unknown to us (i.e., we can only query the filter as a black-box and, for example, cannot invert it). We present two new iterative methods to attack the problem, analyze, and evaluate them on various smoothing and edge-preserving image filters.

scholarly journals Facial Privacy Preservation using FGSM and Universal Perturbation attacks

2022 ◽  
Author(s):  
Nishchal J
Keyword(s):  
Face Recognition ◽  
Privacy Preservation ◽  
Personal Data ◽  
High Accuracy ◽  
Black Box ◽  
Face Image ◽  
Original Image ◽  
User Privacy ◽  
Human Eye ◽  
Face Images

<p>Recent research has established the possibility of deducing soft-biometric attributes such as age, gender and race from an individual’s face image with high accuracy. Many techniques have been proposed to ensure user privacy, such as visible distortions to the images, manipulation of the original image with new face attributes, face swapping etc. Though these techniques achieve the goal of user privacy by fooling face recognition models, they don’t help the user when they want to upload original images without visible distortions or manipulation. The objective of this work is to implement techniques to ensure the privacy of user’s sensitive or personal data in face images by creating minimum pixel level distortions using white-box and black-box perturbation algorithms to fool AI models while maintaining the integrity of the image, so as to appear the same to a human eye.</p><div><br></div>

scholarly journals Explaining Image Classifiers Generating Exemplars and Counter-Exemplars from Latent Representations

2020 ◽  
Vol 34 (09) ◽  
pp. 13665-13668
Author(s):  
Riccardo Guidotti ◽  
Anna Monreale ◽  
Stan Matwin ◽  
Dino Pedreschi
Keyword(s):  
Decision Tree ◽  
Experimental Evaluation ◽  
Decision Rules ◽  
Feature Space ◽  
Saliency Map ◽  
Black Box ◽  
Original Image ◽  
Class A ◽  
Latent Space ◽  
Latent Representations

We present an approach to explain the decisions of black box image classifiers through synthetic exemplar and counter-exemplar learnt in the latent feature space. Our explanation method exploits the latent representations learned through an adversarial autoencoder for generating a synthetic neighborhood of the image for which an explanation is required. A decision tree is trained on a set of images represented in the latent space, and its decision rules are used to generate exemplar images showing how the original image can be modified to stay within its class. Counterfactual rules are used to generate counter-exemplars showing how the original image can “morph” into another class. The explanation also comprehends a saliency map highlighting the areas that contribute to its classification, and areas that push it into another class. A wide and deep experimental evaluation proves that the proposed method outperforms existing explainers in terms of fidelity, relevance, coherence, and stability, besides providing the most useful and interpretable explanations.

Edge-preserving Image Filtering Based on Soft Clustering

2021 ◽  
pp. 1-1
Author(s):  
Yang Yang ◽  
Hongjun Hui ◽  
Lanling Zeng ◽  
Yan Zhao ◽  
Yongzhao Zhan ◽  
...  
Keyword(s):  
Image Filtering ◽  
Edge Preserving ◽  
Soft Clustering

scholarly journals Opening the black box of selection

2019 ◽  
Vol 25 (2) ◽  
pp. 363-382 ◽  
Author(s):  
Sanne Schreurs ◽  
Kitty Cleutjens ◽  
Carlos F. Collares ◽  
Jennifer Cleland ◽  
Mirjam G. A. oude Egbrink
Keyword(s):  
Construct Validity ◽  
Internal Structure ◽  
Selection Procedure ◽  
Black Box ◽  
Situational Judgement Test ◽  
Paradoxical Situation ◽  
Cognitive Diagnostic Modeling ◽  
Personal Competencies ◽  
Set Up ◽  
The Right

Abstract Medical school selection is currently in the paradoxical situation in which selection tools may predict study outcomes, but which constructs are actually doing the predicting is unknown (the ‘black box of selection’). Therefore, our research focused on those constructs, answering the question: do the internal structures of the tests in an outcome-based selection procedure reflect the content that was intended to be measured? Downing’s validity framework was applied to organize evidence for construct validity, focusing on evidence related to content and internal structure. The applied selection procedure was a multi-tool, CanMEDS-based procedure comprised of a video-based situational judgement test (focused on (inter)personal competencies), and a written aptitude test (reflecting a broader array of CanMEDS competencies). First, we examined content-related evidence pertaining to the creation and application of the competency-based selection blueprint and found that the set-up of the selection procedure was a robust, transparent and replicable process. Second, the internal structure of the selection tests was investigated by connecting applicants’ performance on the selection tests to the predetermined blueprint using cognitive diagnostic modeling. The data indicate 89% overlap between the expected and measured constructs. Our results support the notion that the focus placed on creating the right content and following a competency-blueprint was effective in terms of internal structure: most items measured what they were intended to measure. This way of linking a predetermined blueprint to the applicants’ results sheds light into the ‘black box of selection’ and can be used to support the construct validity of selection procedures.

scholarly journals Managing the Adaptation of Modern Society to the Industry 4.0 Based on Information Waves and Impulses

2019 ◽  
Vol 8 (2) ◽  
pp. 438-446 ◽  
Author(s):  
Elena Popkova ◽  
Anastasia Sozinova ◽  
Vera Menshchikova
Keyword(s):  
Internal Structure ◽  
Industry 4.0 ◽  
Wave Interaction ◽  
Modern Society ◽  
Black Box ◽  
Management Process ◽  
Impulse Wave ◽  
The Media ◽  
Four Levels ◽  
Made In

The article analyzes the essence of the management process of adaptation of modern society to the industry 4.0 on the basis of information waves and impulses based on original ideas and scientific discoveries made in the article by A.P. Sukhodolov, I.V. Anokhov and V.A. Marenko Information impulse-wave interaction between the media and society. Scientists proposed an impulsive-oscillating mechanism for the influence of information on four levels of its perception by an individual to explain the essence, logic and internal structure of the management process of modern society's adaptation to the industry 4.0, traditionally represented in the form of an unexplained and little-studied “black box”. As a result, the authors have developed a mechanism for managing the adaptation of modern society to the industry 4.0 on the basis of information waves and pulses.

scholarly journals Geometric Rescaling Algorithms for Submodular Function Minimization

2021 ◽  
Author(s):  
Dan Dadush ◽  
László A. Végh ◽  
Giacomo Zambelli
Keyword(s):  
Iterative Methods ◽  
Polynomial Time ◽  
Black Box ◽  
Submodular Function ◽  
Function Minimization ◽  
Polynomial Algorithms ◽  
Polynomial Time Algorithms ◽  
Wide Range ◽  
Submodular Function Minimization ◽  
Strongly Polynomial

We present a new class of polynomial-time algorithms for submodular function minimization (SFM) as well as a unified framework to obtain strongly polynomial SFM algorithms. Our algorithms are based on simple iterative methods for the minimum-norm problem, such as the conditional gradient and Fujishige–Wolfe algorithms. We exhibit two techniques to turn simple iterative methods into polynomial-time algorithms. First, we adapt the geometric rescaling technique, which has recently gained attention in linear programming, to SFM and obtain a weakly polynomial bound [Formula: see text]. Second, we exhibit a general combinatorial black box approach to turn [Formula: see text]-approximate SFM oracles into strongly polynomial exact SFM algorithms. This framework can be applied to a wide range of combinatorial and continuous algorithms, including pseudo-polynomial ones. In particular, we can obtain strongly polynomial algorithms by a repeated application of the conditional gradient or of the Fujishige–Wolfe algorithm. Combined with the geometric rescaling technique, the black box approach provides an [Formula: see text] algorithm. Finally, we show that one of the techniques we develop in the paper can also be combined with the cutting-plane method of Lee et al., yielding a simplified variant of their [Formula: see text] algorithm.

scholarly journals Hierarchical Guided-Image-Filtering for Efficient Stereo Matching

2019 ◽  
Vol 9 (15) ◽  
pp. 3122 ◽  
Author(s):  
Chengtao Zhu ◽  
Yau-Zen Chang
Keyword(s):  
Stereo Matching ◽  
Image Filtering ◽  
Data Sets ◽  
Matching Accuracy ◽  
Edge Preserving ◽  
Novel Approach ◽  
Combination Scheme ◽  
Image Pairs ◽  
Guided Image Filtering ◽  
The Cost

Stereo matching is complicated by the uneven distribution of textures on the image pairs. We address this problem by applying the edge-preserving guided-Image-filtering (GIF) at different resolutions. In contrast to most multi-scale stereo matching algorithms, parameters of the proposed hierarchical GIF model are in an innovative weighted-combination scheme to generate an improved matching cost volume. Our method draws its strength from exploiting texture in various resolution levels and performing an effective mixture of the derived parameters. This novel approach advances our recently proposed algorithm, the pervasive guided-image-filtering scheme, by equipping it with hierarchical filtering modules, leading to disparity images with more details. The approach ensures as many different-scale patterns as possible to be involved in the cost aggregation and hence improves matching accuracy. The experimental results show that the proposed scheme achieves the best matching accuracy when compared with six well-recognized cutting-edge algorithms using version 3 of the Middlebury stereo evaluation data sets.

Image filtering method using trimmed statistics and edge preserving

2018 ◽  
Vol 12 (7) ◽  
pp. 1086-1094 ◽  
Author(s):  
Weiling Cai ◽  
Ming Yang ◽  
Fengyi Song
Keyword(s):  
Image Filtering ◽  
Edge Preserving ◽  
Filtering Method

scholarly journals Facial Privacy Preservation using FGSM and Universal Perturbation attacks

2022 ◽  
Author(s):  
Nishchal J
Keyword(s):  
Face Recognition ◽  
Privacy Preservation ◽  
Personal Data ◽  
High Accuracy ◽  
Black Box ◽  
Face Image ◽  
Original Image ◽  
User Privacy ◽  
Human Eye ◽  
Face Images

<p>Recent research has established the possibility of deducing soft-biometric attributes such as age, gender and race from an individual’s face image with high accuracy. Many techniques have been proposed to ensure user privacy, such as visible distortions to the images, manipulation of the original image with new face attributes, face swapping etc. Though these techniques achieve the goal of user privacy by fooling face recognition models, they don’t help the user when they want to upload original images without visible distortions or manipulation. The objective of this work is to implement techniques to ensure the privacy of user’s sensitive or personal data in face images by creating minimum pixel level distortions using white-box and black-box perturbation algorithms to fool AI models while maintaining the integrity of the image, so as to appear the same to a human eye.</p><div><br></div>

Sign in / Sign up

Export Citation Format

Share Document