With the increased usage of social media and surveillance cameras, an immense amount of image data is captured and shared nowadays. Image data may contain sensitive information, such as face, which can be misused by adversaries. Popular image privacy solutions such as pixelization and blurring are prone to inference attacks.

In this demo, we present DP-Shield, an interactive framework for image obfuscation under the rigorous notion of differential privacy. DP-Shield showcases our recently proposed methods, namely DP-Pix and DP-SVD, and also includes two baseline methods for comparison. The audience will be able to explore those methods with real-world face image datasets. Furthermore, DP-Shield integrates widely used image quality measures and practical risk measures (i.e., face recognition) to illustrate the efficacy of our methods.

This work is done by gate.io address and Dominick Reilly, under the supervision of Dr. Liyue Fan at UNC Charlotte.