Design of stochastic asymmetric compensation filters for auditory signal processing

This paper introduces a design of asymmetric compensation filters based on stochastic computation for auditory signal processing. Asymmetric compensation filters are used to model gammachirp filters, which well express the performance of human auditory peripheral mechanism and can be used for hearing assisting devices and noise robust speech recognition systems. Using stochastic computation, the asymmetric compensation filters are simply designed using cascaded IIR filters, thanks to a low-complexity implementation of a multiplication. However, the maximum gain of stochastic filters is limited to 1, causing undesirable filter responses. To address the issue, the proposed flexible gain adjusting (FGA) technique normalizes the gain at each IIR filter in stochastic domain while adjusting the total gain in binary domain. For hardware implementation, stochastic asymmetric compensation filters are designed using TSMC 65 nm CMOS technology and the performance is evaluated with the chip layout generated using a standard cell design flow.

[1]  Ralf Schlüter,et al.  Non-stationary feature extraction for automatic speech recognition , 2011, 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[2]  Shie Mannor,et al.  Fully Parallel Stochastic LDPC Decoders , 2008, IEEE Transactions on Signal Processing.

[3]  Brian R Glasberg,et al.  Derivation of auditory filter shapes from notched-noise data , 1990, Hearing Research.

[4]  Roy D. Patterson,et al.  A Dynamic Compressive Gammachirp Auditory Filterbank , 2006, IEEE Transactions on Audio, Speech, and Language Processing.

[5]  Howard C. Card,et al.  Stochastic Neural Computation I: Computational Elements , 2001, IEEE Trans. Computers.

[6]  Vincent C. Gaudet,et al.  Iterative decoding using stochastic computation , 2003 .

[7]  Naoya Onizawa,et al.  VLSI Implementation of Deep Neural Network Using Integral Stochastic Computing , 2017, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[8]  Keshab K. Parhi,et al.  Architectures for IIR digital filters using stochastic computing , 2014, 2014 IEEE International Symposium on Circuits and Systems (ISCAS).

[9]  Naoya Onizawa,et al.  Realization of FIR Digital Filters Based on Stochastic/Binary Hybrid Computation , 2016, 2016 IEEE 46th International Symposium on Multiple-Valued Logic (ISMVL).

[10]  Hermann Ney,et al.  Gammatone Features and Feature Combination for Large Vocabulary Speech Recognition , 2007, 2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07.

[11]  Toshio Irino,et al.  An analysis/synthesis auditory filterbank based on an IIR implementation of the gammachirp , 1999 .

[12]  David J. Lilja,et al.  Using stochastic computing to implement digital image processing algorithms , 2011, 2011 IEEE 29th International Conference on Computer Design (ICCD).

[13]  Naoya Onizawa,et al.  Gammatone filter based on stochastic computation , 2016, 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[14]  Kia Bazargan,et al.  Computation on Stochastic Bit Streams Digital Image Processing Case Studies , 2014, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[15]  Naoya Onizawa,et al.  Area/Energy-Efficient Gammatone Filters Based on Stochastic Computation , 2017, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[16]  Naoya Onizawa,et al.  Gabor Filter Based on Stochastic Computation , 2015, IEEE Signal Processing Letters.

[17]  François W. Primeau,et al.  A million spiking-neuron integrated circuit with a scalable communication network and interface , 2014 .

[18]  Vishwa Gupta,et al.  USE OF MULTIPLE FRONT-ENDS AND I-VECTOR-BASED SPEAKER ADAPTATION FOR ROBUST SPEECH RECOGNITION , 2014 .

[19]  Keshab K. Parhi,et al.  Architectures for Recursive Digital Filters Using Stochastic Computing , 2016, IEEE Transactions on Signal Processing.

[20]  Swathi Priya,et al.  Folded Architecture for Digital Gammatone Filter Used in Speech Processor of Cochlear Implant , 2013 .

[21]  Roy D. Patterson,et al.  A FUNCTIONAL MODEL OF NEURAL ACTIVITY PATTERNS AND AUDITORY IMAGES , 2004 .

[22]  Andrew S. Cassidy,et al.  A million spiking-neuron integrated circuit with a scalable communication network and interface , 2014, Science.

[23]  Yee Whye Teh,et al.  A Fast Learning Algorithm for Deep Belief Nets , 2006, Neural Computation.

[24]  Chutham Sawigun,et al.  Analog complex gammatone filter for cochlear implant channels , 2010, Proceedings of 2010 IEEE International Symposium on Circuits and Systems.

[25]  T. Irino,et al.  A time-domain, level-dependent auditory filter: The gammachirp , 1997 .

[26]  Kia Bazargan,et al.  IIR filters using stochastic arithmetic , 2014, 2014 Design, Automation & Test in Europe Conference & Exhibition (DATE).

[27]  Brian R. Gaines,et al.  Stochastic Computing Systems , 1969 .

[28]  Naoya Onizawa,et al.  High-Accuracy and Area-Efficient Stochastic FIR Digital Filters Based on Hybrid Computation , 2017, IEICE Trans. Inf. Syst..

[29]  Richard F. Lyon,et al.  Practical Gammatone-Like Filters for Auditory Processing , 2006, EURASIP J. Audio Speech Music. Process..

[30]  T. Irino,et al.  A compressive gammachirp auditory filter for both physiological and psychophysical data. , 2001, The Journal of the Acoustical Society of America.