Automatic Speaker Recognition (ASR)

A computational method for forensic speaker comparison using i-vector or x-vector feature extraction from audio recordings to compute a likelihood ratio expressing the probability of the questioned recording given the hypothesis that the same speaker produced both samples.