Abstract. Maximum distance separable (MDS) matrices have applications not only in coding theory but also are of great importance in the design of block ciphers and hash functions. It is highly nontrivial to find MDS matrices which could be used in lightweight cryptography. In a crypto 2011 paper, Guo et. al. We also propose more generic constructions of MDS matrices e.g. we construct lightweight 4 × 4 and 5 × 5 MDS matrices over F2n for all n ≥ 4. An algorithm is presented to check if a given matrix is MDS. The algorithm follows from the basic properties of MDS matrix and is easy to implement.

No abstract

No abstract

This paper proposes an efficient algorithm to perform a privacy-preserving speaker verification based on the iVector and linear discriminant analysis. In this research we have considered a scenario in which the users enrol their voice biometric with the third-party service providers to access the different services (i.e., banking). Once the enrolment is completed, the users can verify themselves to the system using their voice instead of passwords. Since the voice is unique for everyone, storing the extracted voice features of the user at the third-party server raises several privacy concerns. To address this challenge, this paper proposes a novel technique based on randomisation to perform voice authentication, which allows the user to enrol and verify their voice in the randomised domain. To achieve this, the iVector based speaker verification technique has been redesigned to work on the randomised domain. The proposed algorithm is validated using the TIMIT dataset. In addition, the proposed algorithm does not compromise the accuracy due to the randomisation as the additional complexity due to the randomisation is negligible.

This is the accepted version of the paper.This version of the publication may differ from the final published version. Permanent ABSTRACTThis paper presents a strategy for enabling speech recognition to be performed in the cloud whilst preserving the privacy of users. The approach advocates a demarcation of responsibilities between the client and server-side components for performing the speech recognition task. On the client-side resides the acoustic model, which symbolically encodes the audio and encrypts the data before uploading to the server. The server-side then employs searchable encryption to enable the phonetic search of the speech content. Some preliminary results for speech encoding and searchable encryption are presented.