DocumentCode :
626781
Title :
A two-class information concealing system based on compressed sensing
Author :
Cambareri, Valerio ; Haboba, Javier ; Pareschi, Fabio ; Rovatti, H. Riccardo ; Setti, Gianluca ; Kwok-Wo Wong
Author_Institution :
DEI - Univ. of Bologna, Bologna, Italy
fYear :
2013
fDate :
19-23 May 2013
Firstpage :
1356
Lastpage :
1359
Abstract :
We elaborate on the possibility of exploiting the (pseudo)random projection operator, which is at the heart of the most common architecture for compressed sensing, to prevent access to the acquired information by unauthorized receivers. In low-resource applications, this approach may make dedicated cryptographic layers unnecessary when the security requirement is not particularly high. Beyond proving that the proposed system is at least asymptotically immune to straightforward statistical attacks, we also exploit the sensitivity of compressed sensing recovery algorithms to the complete knowledge of the projection matrix to introduce two-class protection. The encoding is such that first-class decoders can retrieve the signal to its full resolution while second-class decoders are able to retrieve only a degraded version of the same signal. Examples are given with reference to ECG signal acquisition.
Keywords :
compressed sensing; decoding; electrocardiography; medical signal detection; security of data; ECG signal acquisition; compressed sensing recovery algorithms; dedicated cryptographic layers; first-class decoders; pseudorandom projection operator; second-class decoders; security requirement; straightforward statistical attacks; two-class information concealing system; two-class protection; unauthorized receivers; Compressed sensing; Decoding; Electrocardiography; Encoding; Noise; Receivers; Security;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Circuits and Systems (ISCAS), 2013 IEEE International Symposium on
Conference_Location :
Beijing
ISSN :
0271-4302
Print_ISBN :
978-1-4673-5760-9
Type :
conf
DOI :
10.1109/ISCAS.2013.6572106
Filename :
6572106
Link To Document :
بازگشت