An efficient lattice-based phonetic search method for accelerating keyword spotting in large speech databases

@article{Tetariy2013AnEL,
  title={An efficient lattice-based phonetic search method for accelerating keyword spotting in large speech databases},
  author={Ella Tetariy and Michal Gishri and Baruch Har-Lev and Vered Aharonson and Ami Moyal},
  journal={International Journal of Speech Technology},
  year={2013},
  volume={16},
  pages={161-169}
}
This paper describes an algorithm for the reduction of computational complexity in phonetic search KeyWord Spotting (KWS). This reduction is particularly important when searching for keywords within very large speech databases and aiming for rapid response time. The suggested algorithm consists of an anchor-based phoneme search that reduces the search space by generating hypotheses only around phonemes recognized with high reliability. Three databases have been used for the evaluation: IBM… CONTINUE READING

Figures, Tables, Results, and Topics from this paper.

Key Quantitative Results

  • The results indicate that using the anchor-based approach in KWS reduces the size of search space and the processing time by 86–90 % in comparison to an exhaustive search, while at the same time showing a 31 % reduction in the FAR for the VM database and a 50 % reduction in the FAR for Mac.
  • We demonstrated a reduction of almost 90 % in search space and computational complexity of phonetic search KWS by using a phoneme anchor point algorithm.
  • The results showed that the KWS mechanism produces a 100 % detection rate with a relatively low false alarm rate.
  • Importantly, we were able to reach the main goal of this research, which was to present an efficient Phonetic Search KWS method with a 90 % reduction in search space and consequently runtime, while even improving the FAR in comparison to the exhaustive search with some decrease in the detection rate.

Citations

Publications citing this paper.
SHOWING 1-5 OF 5 CITATIONS

A fast approach to spoken term detection based on prosodic dynamic features

  • 2015 IEEE International Conference on Progress in Informatics and Computing (PIC)
  • 2015

Design of Manipuri Keywords Spotting System using HMM

  • 2013 Fourth National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics (NCVPRIPG)
  • 2013
VIEW 1 EXCERPT

Phonetic Search Methods for Large Speech Databases

  • Springer Briefs in Electrical and Computer Engineering
  • 2013
VIEW 1 EXCERPT
CITES BACKGROUND

Cross-language phoneme recognition for under-resourced languages

  • 2012 IEEE 27th Convention of Electrical and Electronics Engineers in Israel
  • 2012

References

Publications referenced by this paper.
SHOWING 1-10 OF 16 REFERENCES

Phonetic search using an anchor-based algorithm

  • 2010 IEEE 26-th Convention of Electrical and Electronics Engineers in Israel
  • 2010
VIEW 2 EXCERPTS

Dynamic match phone-lattice searches for very fast and accurate unrestricted vocabulary keyword spotting

  • Proceedings. (ICASSP '05). IEEE International Conference on Acoustics, Speech, and Signal Processing, 2005.
  • 2005
VIEW 1 EXCERPT

Voicemail Corpus Part II. Philadelphia, USA: Linguistic Data Consortium (LDC)

M. Padmanabhan, B. Kingsbury, +3 authors G Saon
  • 2002
VIEW 2 EXCERPTS