A time-frequency aware cochlear implant: algorithm and system

  • Authors:

  • Songping Mai;Yixin Zhao;Chun Zhang;Zhihua Wang

  • Affiliations:

  • Lab. of Integrated Circuits and Systems Design, Graduate School at Shenzhen, Tsinghua University, Shenzhen, China;Lab. of Integrated Circuits and Systems Design, Graduate School at Shenzhen, Tsinghua University, Shenzhen, China;Lab. of Integrated Circuits and Systems Design, Graduate School at Shenzhen, Tsinghua University, Shenzhen, China;Lab. of Integrated Circuits and Systems Design, Graduate School at Shenzhen, Tsinghua University, Shenzhen, China

  • Venue:
  • ISNN'12 Proceedings of the 9th international conference on Advances in Neural Networks - Volume Part II
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

A time-frequency aware (TFA) cochlear implant system and its speech processing strategy (TFA-CIS) are presented in this paper. The system is built upon an implanted low-power digital signal processor (THUCIDSP). And the TFA-CIS strategy takes advantages of the time-frequency aware wavelet packet (WP) filter bank (FB) and WP envelope detector. The joint use of the DSP hardware and the TFA strategy gives birth to a very low power cochlear implant system. Implementation result shows that the complexity of the new TFA-CIS algorithm is 1.82 million instructions per second (MIPS) with db1 wavelet base, and the power dissipation of the DSP is 2.11mW @1.8V, 3MHz.