SRG - Speech Recognition Group

The Speech Recognition group covers most aspects of the speech, speaker and language recognition: signal processing, acoustic, phonetic and lexical modeling.

Most relevant scientific contributions: large vocabulary automatic speech recognition, speaker and language recognition systems robust to channel variations.

Most relevant applications in the past years:

  • Collaboration to the development of the Loquendo ASR speech recognizer
  • Collaboration to the development of an automatic Directory Assistance systems for Telecom Italia
  • Collaboration to the development of the Nuance speaker recognition system

Skills

  • Pattern recognition
  • Machine Learning
  • Parallel  Computation
  • Speech and Language Processing

Projects and publications

  • Selected funded research projects
    • SPEAKER RECOGNITION AND DIARIZATION
      CUMANI SANDRO
      2021 - (Current)
    • SPEAKER RECOGNITION AND DIARIZATION
      CUMANI SANDRO
      2022 - 2022 (Completed)
    • SPEAKER RECOGNITION AND DIARIZATION
      CUMANI SANDRO
      2020 - 2020 (Completed)
    • SPEAKER RECOGNITION USING DNNS AND HMMS
      CUMANI SANDRO
      2018 - 2018 (Completed)