Differentially Private Image Classification Using Support Vector Machine and Differential Privacy

Makhamisa Senekane

doi:10.3390/make1010029

Abstract

The ubiquity of data, including multi-media data such as images, enables easy mining and analysis of such data. However, such an analysis might involve the use of sensitive data such as medical records (including radiological images) and financial records. Privacy-preserving machine learning is an approach that is aimed at the analysis of such data in such a way that privacy is not compromised. There are various privacy-preserving data analysis approaches such as k-anonymity, l-diversity, t-closeness and Differential Privacy (DP). Currently, DP is a golden standard of privacy-preserving data analysis due to its robustness against background knowledge attacks. In this paper, we report a scheme for privacy-preserving image classification using Support Vector Machine (SVM) and DP. SVM is chosen as a classification algorithm because unlike variants of artificial neural networks, it converges to a global optimum. SVM kernels used are linear and Radial Basis Function (RBF), while ϵ -differential privacy was the DP framework used. The proposed scheme achieved an accuracy of up to 98%. The results obtained underline the utility of using SVM and DP for privacy-preserving image classification.

Highlights

With a massive increase in the collection and storage of personal data such as medical records, financial records and census data and web search histories, the concern for privacy has been exacerbated [1]
It can be observed that the performance of privacy-preserving image classification Support Vector Machine (SVM) is comparable to that of the pure image classification SVM
These results demonstrate that the performance of pure SVM is comparable to that of Differential Privacy (DP)-SVM, since their misclassification errors are comparable

Summary

Introduction

With a massive increase in the collection and storage of personal data such as medical records, financial records and census data and web search histories, the concern for privacy has been exacerbated [1]. Unlike anonymization schemes discussed above, DP provides information-theoretic guarantee that the participation of the individual(s) in a statistical database would not be revealed It has since become a gold standard of privacy-preserving data analysis. We propose a privacy-preserving image classification scheme using support vector machine (SVM) and DP. The key contribution of the work reported in this paper is the design of privacy-preserving image classification algorithm which makes use of SVM and -DP. We provide background information on SVM, DP and privacy-preserving machine learning, as well as related work concerning privacy-preserving machine learning

Support Vector Machine

Differential Privacy

Privacy-Preserving Machine Learning

Privacy-Preserving Image Classification Algorithm

Results and Discussion

Conclusions