Adversarial Privacy-preserving Filter

Zhang, Jiaming; Sang, Jitao; Xian, Zhao; Huang, Xiaowen; Sun, Yanfeng; Hu, Yongli

doi:10.1145/3394171.3413906

Cited by 33 publications

(22 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The filter integrates universal adversarial perturbation with image-specific adversarial perturbation to generate adversarial images, which advances the performance on portrait privacypreserving. More examples of face recognition adversarial attacks by this filter are discussed in [8].…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Apf

Xian

Zhang

Huang

2021

Proceedings of the 29th ACM International Conference on Multimedia

Self Cite

View full text Add to dashboard Cite

While widely adopted in practical applications, face recognition has been disputed on the malicious use of face images and potential privacy issues. Online photo sharing services accidentally act as the main approach for the malicious crawlers to exploit face recognition to access portrait privacy. In this demo, we propose an adversarial privacy-preserving filter, which can preserve face image from malicious face recognition algorithms. This filter is generated by an end-cloud collaborated adversarial attack framework consisting of three modules: (1) Image-specific gradient generation module, to extract image-specific gradient in the user end; (2) Adversarial gradient transfer module, to fine-tune the image-specific gradient in the server; and (3) Universal adversarial perturbation enhancement module, to append image-independent perturbation to derive the final adversarial perturbation. A short video about our system is available at here. CCS CONCEPTS• Security and privacy → Privacy protections; Usability in security and privacy.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Compared with the traditional adversarial attack algorithms, our method achieves better performance on portrait privacypreserving without affecting the image quality. More experimental results are illustrated in [8].…”

Section: Universal Perturbation Enhancement Modulementioning

confidence: 99%

Apf

Xian

Zhang

Huang

2021

Proceedings of the 29th ACM International Conference on Multimedia

Self Cite

View full text Add to dashboard Cite

show abstract

“…In medical imaging darknet53 used for detection of covid-19 [ 34 ] computed aided covid-19 detection [ 36 ] for MRI scan brain tumor data augmentation [ 35 ] YOLO V3 has been used to identify red lesions in retinal fundus images [ 37 ]. smart medical autonomous distributed system for diagnosis [ 38 ], melanoma detection [ 39 ].…”

Section: Methodsmentioning

confidence: 99%

“…Contractive auto-encoder (CAE) deep neural networks work as a robust model against adversarial examples with high accuracy [36]. Some applications of adversarial attacks using pre-trained deep learning (DL) models in computer vision tasks include, e.g., visual classification [37], textual data system [38], privacy-preserving filter [39], object detector [40], image segmentation [41], natural language processing [42], data fusion [43], hybrid digital watermarking and text document retrieval [44], fingerprint liveness detection [45], person re-identification [46], time series classification [47], human activity recognition [48], face recognition [49], handwritten signature verification [50], and multi-objective reinforcement learning [51].…”

Section: Introductionmentioning

confidence: 99%

Adversarial Attack and Defence through Adversarial Training and Feature Fusion for Diabetic Retinopathy Recognition

Lal

Rehman

Shah

et al. 2021

Sensors

View full text Add to dashboard Cite

Due to the rapid growth in artificial intelligence (AI) and deep learning (DL) approaches, the security and robustness of the deployed algorithms need to be guaranteed. The security susceptibility of the DL algorithms to adversarial examples has been widely acknowledged. The artificially created examples will lead to different instances negatively identified by the DL models that are humanly considered benign. Practical application in actual physical scenarios with adversarial threats shows their features. Thus, adversarial attacks and defense, including machine learning and its reliability, have drawn growing interest and, in recent years, has been a hot topic of research. We introduce a framework that provides a defensive model against the adversarial speckle-noise attack, the adversarial training, and a feature fusion strategy, which preserves the classification with correct labelling. We evaluate and analyze the adversarial attacks and defenses on the retinal fundus images for the Diabetic Retinopathy recognition problem, which is considered a state-of-the-art endeavor. Results obtained on the retinal fundus images, which are prone to adversarial attacks, are 99% accurate and prove that the proposed defensive model is robust.

show abstract

“…A different type called Universal Adversarial Perturbation (UAP), however, is sample-independent which misleads most of the input samples for given model and category [31]. Therefore, UAP can serve as probe to explore what features the model relies on [36,49]. In this subsection and the subsequent analysis, we will employ UAP as tool to analyze the representative information used in machine learning algorithms (mainly non-semantic).…”

Section: Justification Analysismentioning

confidence: 99%

Benign Adversarial Attack: Tricking Algorithm for Goodness

Sang¹,

Zhao²,

Zhang³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

In spite of the successful application in many fields, machine learning algorithms today suffer from notorious problems like vulnerability to adversarial examples. Beyond falling into the cat-and-mouse game between adversarial attack and defense, this paper provides alternative perspective to consider adversarial example and explore whether we can exploit it in benign applications. We first propose a novel taxonomy of visual information along task-relevance and semantic-orientation. The emergence of adversarial example is attributed to algorithm's utilization of task-relevant non-semantic information. While largely ignored in classical machine learning mechanisms, task-relevant non-semantic information enjoys three interesting characteristics as (1) exclusive to algorithm, (2) reflecting common weakness, and (3) utilizable as features. Inspired by this, we present brave new idea called benign adversarial attack to exploit adversarial examples for goodness in three directions: (1) adversarial Turing test, (2) rejecting malicious algorithm, and (3) adversarial data augmentation. Each direction is positioned with motivation elaboration, justification analysis and prototype applications to showcase its potential. CCS CONCEPTS• Computing methodologies → Computer vision; Computer vision problems.

show abstract

Adversarial Privacy-preserving Filter

Cited by 33 publications

References 31 publications

Apf

Apf

Adversarial Attack and Defence through Adversarial Training and Feature Fusion for Diabetic Retinopathy Recognition

Benign Adversarial Attack: Tricking Algorithm for Goodness

Contact Info

Product

Resources

About