Robust and efficient hashing framework for industrial surveillance

The recent advent in the multimedia tools and their wide availability lead to a critical issue of protecting the privacy of multimedia content in cyber-physical security of industrial set-ups predominantly in surveillance. This paper emphasizes on preserving the authenticity of multimedia content in industrial surveillance by presenting an efficient hashing technique based on normalization, log-polar mapping and singular value decomposition. The core idea is to produce a hash sequence from the key-frames extracted from industrial surveillance video providing better robustness and security. For this purpose, the input key-frame is first normalized to make it resilient against the affine distortions. Log-polar mapping is then applied on the normalized key-frame, and an initial hash sequence is generated using the properties of singular value decomposition. At last, a randomization process is applied to construct the final hash sequence. Extensive experiments on various key-frames are conducted to demonstrate the robustness of the proposed framework against various intentional/unintentional distortions. © 2022, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.