Machine Intelligence for Speech Audio and Multimedia (MISAM) research group focuses on intelligent processing of speech audio and multimedia using machine learning and signal processing techniques.

Research in our group includes the following topics

  • Deep neural networks for audio/speech data
  • Speech enhancement
  • Audio source separation
  • Noise-robust speech recognition
  • Secure biometrics
  • Sparse signal recovery

Our recent research involves using deep neural networks and deep recurrent neural networks for audio and multimedia data. Deep learning is becoming very important in solving pattern recognition and signal processing problems.

Research on noise-robust speech processing and recognition is booming recently. Especially, companies like Google, Amazon, Microsoft, Facebook are hiring new researchers in the field of speech recognition using deep learning.

We have research collaborations with Mitsubishi Electric Research Laboratories in Cambridge, MA. We jointly work on speech enhancement, source separation, deep learning, noise-robust speech recognition.

Please contact Prof. Hakan Erdogan if you would like to receive more information about research in the MISAM lab.