Environment recognition for digital audio forensics using MPEG-7 features and separate modeling technique

Environment recognition from digital audio for forensic application is a growing area of interest. However, compared to other branches of audio forensics, it is a less researched one. Especially less attention has been given to detect environment from files where foreground speech is present, which is a forensic scenario. In this paper, we perform several experiments focusing on the problems of environment recognition from audio particularly for forensic application. Experimental results show that the task is easier when audio files contain only environmental sound than when they contain both foreground speech and background environment. We propose a full set of MPEG-7 audio features and separate models for each environment sound and human speech to improve the accuracy. In the experiments, the proposed approach significantly increases the recognition accuracy of environment sound even in presence of high amount of foreground human speech