文摘
Source cell-phone identification has become a hot topic in multimedia forensics recently. In this paper, we propose a novel cell-phone identification method based on the recorded speech files. Device self-noise is considered as the fingerprint of the cell-phone, and the self-noise is estimated from the near-silent segments of recording. Moreover, two categories of spectral features of self-noise, i.e., spectral shape features (SN-SSF) and spectral distribution features (SN-SDF), are extracted for closed-set classification using SVM classifier. Experimental results show that the self-noise has the ability to identify the cell-phones of 24 different models, and identification accuracies of 89.23% and 94.53% have been obtained for SN-SSF and SN-SDF, respectively. To the best of our knowledge, it is the first attempt to comprehensively define the self-noise of cell-phone and furthermore apply it to source identification issue of audio forensics.