Sparse representation and compressed sensing
Signal decomposition techniques
Big data analytics and machine learning
Audio and speech processing
Biomedical signal and image analysis
Above techniques are applied in:
Audio scene analytics
Atrial fibrillation and ventricular fibrillation
Sleep signal processing
- Defining new ways of robust
signal feature extraction that incorporates signal non-stationarity,
and classifying using simple pattern recognizers. This has numerous
applications in various fields including the emerging areas of
bioinformatics and biometrics.
- Developing a new paradigm for
data hiding using true time-frequency tools. Including the
competency of the chirp detectors to perform as forward error
correction codes. There is a compelling anxiety among the peers to
see how the chirp detectors perform against the well established FEC
schemes used in digital transmission.
- Investigating non-conventional
data compression schemes (time-frequency, wavelet-based, and grammar
codes) for images and audio files. Based on the experiences gained
in this field SAR members will be embarking on network-centric
coding schemes with inbuilt information integrity and security.
- Fine tuning established signal
enhancement schemes for ground vehicle tracking, improved listening
experience for hearing impaired people, and in determining better
statistical parameters for lightening protection applications.
- Working with real world data
samples: knee sounds, pathological voice, EEG from alcoholics, EMG
data from children's hospital, digital mammograms, GPS signals
acquired in downtown Toronto, computer keystroke data, ultrasound
signals from cells cultivated in lab, pulse (blood) volume signals,
lightening data from CN Tower, and commonly used
speech/audio/image/sports video data.