An efficient retrieval algorithm of encrypted speech based on inverse fast Fouriertransform and measurement matrix

Abstract

In this paper, we present an efficient retrieval algorithm for encrypted speech based on an inverse fast Fourier transform and measurement matrix. Our approach improves query performance, as well as retrieval efficiency and accuracy, compared to existing content-based encrypted speech retrieval methods. Our proposed algorithm constructs a perceptual hash scheme using perceptual hash sequences from original speech files. By classifying the sequences and applying run-length compression, we decrease the cloud storage required for the hash index. We secure the speech database by encrypting it with Henon chaos scrambling, which offers excellent resistance to attacks. Experimental results show that the robustness, discrimination, and feature extraction efficiency of our proposed method are better than the existing alternatives, with good recall and precision ratios and with high retrieval efficiency and accuracy.

Encrypted speech retrieval based on long sequence Biohashing

“…When τ = 0.16, the number of misjudgments for each 1.0 × 10 73 speech segments is 6.6488. Under the same conditions, it is 3.8 × 10 43 times smaller than [34], 4.9 × 10 43 times smaller than [35], 5.9 × 10 57 times smaller than [37], 1.0 × 10 59 times smaller [25].…”

“…For encrypted speech, how to improve the security is the key, i.e., the disclosure of plaintext data. Chaos encryption [25,32,35], digital watermark encryption [7,37], ESRDH encryption [2], DCT encryption [36] and m-sequence encryption [34] can improve the security of plaintext data by changing the original speech order, embedding and changing the original speech data. In recent years, with the emergence of a large number of feature extraction methods (i.e., partially supervised depth hash [11], deep cross-modal learning [6,30], deep learning [20], wavelet transform [23]), the retrieval accuracy of speech retrieval system has been greatly improved.…”