Reconstruction of Privacy-Sensitive Data from Protected Templates

ABSTRACT The problem of template privacy protection in view of reconstruction attacks is well recognized and various generic In this paper, we address the problem of data reconstruction methods were proposed such as fuzzy commitment schemes from privacy-protected templates, based on recent [3] and secure sketches [4], helper data based methods [5, concept of sparse ternary coding with ambiguization (STCA). We do not pretend to be exhaustive in our overview addition of ambiguization noise to satisfy the privacy-utility and refer interesting readers to [8]. The theoretical privacy-preserving the STCA was proposed that combines and extends the encoding properties of STCA have been validated on synthetic data. The STCA ensures the protection threats linked to reconstruction based on recent deep of both templates and queries in authentication and identification reconstruction algorithms are still open problems. Our results systems against the adversarial reconstruction and demonstrate that STCA still achieves the claimed theoretical clustering [9, 10].