Detecting Replicas within Audio Evidence Using an Adaptive Audio Fingerprinting Scheme

摘要

In this paper we propose a passive audio authentication method based on an adaptive audio fingerprinting scheme to detect forgeries produced by the replication of an audio interval within the same evidence. Several audio fingerprinting systems are analyzed, and an adaptive binary audio fingerprinting scheme based on the Fourier spectrum distribution is proposed. The adaptive system is theoretically and empirically adjusted for each audio evidence to detect short replicas. Simulations are performed to analyze the robustness against time, frequency-domain, and compression distortions. The power to discriminate repeated text speech and distinguish it from audio replicas is also analyzed.