1 / 25

My research strands in MSA

My research strands in MSA. Barry Cheetham Review of past year Plans for future. *Derek Molyneux *Meau Shin Ho Ning Li Li-rong He Lizzie X Ni. Yen-ha Hiew Abdul Eshhubi Andrew Earis Tolis Tramantzas. Personnel. Research themes. Speech coding (3) Patient monitoring (1)

cleaver
Télécharger la présentation

My research strands in MSA

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. My research strands in MSA Barry Cheetham Review of past year Plans for future

  2. *Derek Molyneux *Meau Shin Ho Ning Li Li-rong He Lizzie X Ni Yen-ha Hiew Abdul Eshhubi Andrew Earis Tolis Tramantzas Personnel

  3. Research themes • Speech coding (3) • Patient monitoring (1) • Implementation (1) • Distance learning networks (2) • Music analysis (1) • Speech over WLAN (1)

  4. Speech coding for mobile & VoIP • Discrete all-pole modelling applied to speech coding. • Sinusoidal transform coding of speech using DAP. • Variable bit-rate coding using DAP-STC.

  5. Speech personnel

  6. Speech personnel

  7. DAP-STC • 2G GSM uses CELP at 13kb/s+FEC • 3G & IP need AMR down to 2.4kb/s • “Toll” quality • CELP fails at bit-rates below 4kb/s • Use sinusoidal / interpolative coding • DAP more accurate than LPC.

  8. DAP-STC fv fv fv DAP spectral envelope Freq’ncies Compute magntude spectrum Peak-picking Speech input X DFT LSP Amplitudes Window Pitch frequency Pitch & voicing detection Voicing probability (Pv)

  9. Patient monitoring

  10. Monitoring of chronic cough • Collaboration with Wythenshawe & Fazakerley Hospitals. • Recognition of cough waveforms • Elimination of speech/ background • 24 hour monitoring /data storage • GPRS radio link

  11. Cough monitoring(cont) • Speech recognition techniques • Statistical measurements • Participation in trials • Certain medical procedures may be proved misleading

  12. Expulsive phase Mid phase Final burst A COUGH

  13. Low power implementation of DSP for mobile terminals. • Lizzie is here!!

  14. Retargetable compiler for low power on CADRE • Motivation: • New asynchronous DSP core—CADRE • Retargetable compiler support • Speed, code size & power efficiency • Techniques at software level • Research Objectives: • Methodology for C compilation flow for target. • Results so far • Shown to reduce switching activity at circuit level.

  15. Distance learning networks (LiRong) • Real time over fixed or wireless networks • Flow & congestn control for speech/image/data • Based on measurements of delay, jitter, lost packet probability • New ‘RTMTCP’ for distance learning • Dynamic priority assignment • Adapted to on-line tutorials

  16. Distance learning by Internet • Tolis is here!

  17. Musical Analysis: Andrew Earis

  18. Expressive musical performance analysis • Measure intentional variations in the way notes played. • Timing, dynamics, tone quality, vibrato, etc. • Extract from CD with ref to the score in midi form. • Display to augment the score. • Many applications, e.g. practising, analysis, compression • Classical time-frequency analysis problem.

  19. Voice over WLAN • Integrating DECT PABX with OFDM based IEEE802.11 WLAN • Use single band, e.g. 2.4 GHz • Share resources, simplify maint’nce • Modify DECT radio interface & packetise the speech • Operate in ‘contention mode’

  20. Wireless LAN Cordless phone Laptop AP Cordless phone Laptop PDA LAN connection

  21. Principle: OFDM & single carrier

  22. The future • Ning’s & Li-Rong’s theses • Publish DAP-STC work in full • Grant application for cough work • Grants for VoWLAN / Music • Framework 6

  23. Opportunities Framework 6 Collaboration with ASCOM Spectrum Efficiency call (RCA)

  24. Goals for next year • One new student for OFDM work • Two successful PhDs & 1 MSc • At least 1 grant. • 2-3 journal papers + conf.

  25. Publications this year • Two IEEE ICASSP conf. • Two other conf. (Andrew) • One IEE colloquium paper (YH) • 4 medical conf. • COST presentation

More Related