Cellphone identification using noise estimates from recorded audio

Rapid developments in technologies related to cell phones have resulted in their much broader usage than mere talking devices used for making and receiving phone calls. User-generated audio recordings from cell phones can be very helpful in a number of forensic applications. This paper proposes a novel system for cellphone identification from speech samples recorded using the cellphone. The proposed system uses features based on estimates of noise associated with recordings and classifies them using sequential minimal optimization (SMO) based Support vector machine (SVM). The performance of the proposed system is tested on a custom database of twenty-six cell phones of five different manufacturers. The proposed system shows promising results with average classification accuracy around 90% for classifying cell phones belonging to five different manufacturers. The average classification accuracy reduces when all the cell phones belong to the same manufacturer.