We are able to recognize our family members and friends solely by their voice. This is possible due to the speaker cues in nonlinguistic information that speech carries simultaneously with the linguistic message. This paper gives a partial review of the human speaker recognition literature in terms of the definitions, the applied methods, and the results. The latest experimental method, fMRI, opened a new avenue in the research of this cognitive process, thus the author focuses on studies employing this technique. He compares their results with those of behavioral studies and highlight their implications for speech technology. He also makes an attempt to summarize the sporadically published psychological models of speaker recognition.