Academy & Industry Research Collaboration Center (AIRCC)

Volume 10, Number 02, March 2020

Privacy-preserving Pattern Recognition with Image Compression


Takayuki Nakachi1 and Hitoshi Kiya2, 1Nippon Telegraph and Telephone Corporation, Japan and 2Tokyo Metropolitan University, Japan


In this paper, we propose a privacy-preserving pattern recognition scheme that well supports image com- pression. The proposed scheme is based on secure sparse coding using a random unitary transform. It offers the following two prominent features: 1) It is capable of pattern recognition in the encrypted image domain. Even if data leaks, privacy can be maintained because data remains encrypted. 2) It realizes Encryption-then-Compression (EtC) systems, where image encryption is conducted prior to compression. The pattern recognition can be carried out in the compressed signal domain using a few sparse coefficients. Based on the pattern recognition result, it can compress the selected images with high quality by estimat- ing sufficient number of sparse coefficients. We use the INRIA dataset to demonstrate its performance in detecting humans. The proposal is shown to realize human detection with encrypted images and efficiently compress the images selected in the image recognition stage.


Surveillance Camera, Pattern Recognition, Secure Computation, Sparse Coding, Random Unitary Trans- form