Avoice: Speaker Identification System

In addition to concepts transferred by human speech, it contains information about the language used, the speaker’s emotional state, gender and identity. The goal of a speaker recognition system (verification/identification) is to extract information from speech signal and to recognize who the speaker is. Speaker recognition encompasses speaker identification and speaker verification applications. In a speaker identification system, the person is identified from his/her speech and in a speaker verification system, the speaker claims to be a known person and the systems verifies whether the claim is true or not.

Avoice is a speaker identification system. It can be used in various security and access control tasks, alone or together with other security methods. Avoice speaker identification engine operates under various constraints and that can run online or offline. In addition, the identification of speaker can be performed remotely (e.g. on the phone).


Avoice Applications:

  • Search and identification of criminals using their voice
  • Forensic identification applications


Avoice Features:

  • High accuracy
  • High-speed processing
  • Usable for online or offline applications
  • Ability to manage possible target speakers by the operator
  • Online monitoring of multiple radio, TV, or other multimedia channels
  • Support a variety of voice formats and communication channels such as telephone, mobile, VOIP, …


For more information about system requirements and installation guidelines of Avoice, please contact the company.