← Registry / Multimodal Conversation

AIM

Speaker Identity Recognition

MMC-SIR-V2.5

Download definition

Submissions for this AIM must match this definition: the same ports, API profile and internal wiring.

Description

This AIM implements the speaker recognition function: it identifies a speaker based on the Speech Object that contains their speech. It is the audio counterpart of Face Identity Recognition.

Inputs and outputs 5 in · 1 out

Inputs

  • AuxiliaryText optional OSD-TXO-V1.5
  • BasicSpeechObject OSD-BSO-V1.5 or OSD-SPO-V1.5
  • SpeechOverlap optional MMC-SOL-V2.5
  • SpeechSceneGeometry optional OSD-SSG-V1.5
  • AudioSceneDescriptors optional OSD-BAS-V1.5

Outputs

  • SpeakerID OSD-IID-V1.5

Used in 1 AIM

Documentation

Implementations 0

None published yet.