Comparison of Different Wavelet Sub-Band Features in the Classification of Indonesian Stop Consonants in CV Syllable Context

  • Abstract
  • Keywords
  • References
  • PDF
  • Abstract

    In the research field of signal processing by using wavelet method there are some factors affecting the accuracy of recognition such as the selection of the sub-band parameter, the selection of suitable mother wavelet or coefficient, and the determination of decomposition level. This paper presents a comparative study of three wavelet-based sub-bands (WBSB) combined with the moving average energy (MAE) features for classification of Indonesian stop consonants in consonant-vowel (CV) context. Three different feature sets used in this study are the MAE of each different wavelet sub-band using mother wavelet of daubechies2. The first feature set is the MAE taken from standard wavelet packet (WP) sub-band at the 4th level of decomposition denoted as WBSB. Whereas the second and third feature sets are the MAE taken from WP which the sub-band is selected based on the previous research denoted as WBSB1 and WBSB2. For the classification of the stops sound signal after feature extraction process, two different classifiers were used, based on multi-layer perceptron and random forest. The experimental result showed that the performance rank of feature extraction method were WBSB, WBSB1, and WBSB2, respectively.



  • Keywords

    Feature extraction; frequency sub-band; stop consonants; wavelet.

  • References

      [1] Boccaletti S, Giaquinta A, and Arecchi FT (1997), “Adaptive recognition and filtering of noise using wavelets,” Physical Review Journal. Rev. E, vol. 55, no. 5, pp. 5393–5397. DOI: 10.1103/PhysRevE.55.5393.

      [2] Holschneider M (1988), “On the wavelet transforms of fractal objects,”Journal of Statistical Physics, vol. 50, pp. 953–993. DOI: 10.1007/BF01019149.

      [3] Sharma RP, Farooq O, and Khan I (2013), “Wavelet based sub-band parameters for classification of unaspirated Hindi stop consonants in initial position of CV syllables,” International Journal of Speech Technology, vol. 16, no. 3, pp. 323–332. DOI: 10.1007/s10772-012-9185-x.

      [4] Hardjono FL and Fox RA (2011), “Stop Consonant Characteristics: VOT and Voicing in American-Born-Indonesian Children’s Stop Consonants,” The Ohio State University.

      [5] Hasan A and Dardjowidjojo S (2003), Tata Bahasa Baku Bahasa Indonesia (Indonesian Grammar), Vol.3. Jakarta: Balai Pustaka.

      [6] Amalia N, Fahrudi AE, and Nasrulloh AV (2013), “Indonesian Vowel Recognition using Artificial Neural Network based on the Wavelet Features,” International Journal Electronic and Computer Engineering, vol. 3, no. 2, pp. 260–269.

      [7] Hidayat R, Kristomo D, and Togarma I, “Feature extraction of the Indonesian phonemes using discrete wavelet and wavelet packet transform,” in 2016 8th International Conference on Information Technology and Electrical Engineering (ICITEE), 2016, pp. 478–483. DOI: 10.1109/ICITEED.2016.7863310.

      [8] S. Hidayat, R. Hidayat, and T. B. Adji, “Speech recognition of CV-patterned indonesian syllable using MFCC, wavelet and HMM,” J. Ilm. Kursor, vol. 8, no. 2, pp. 67–78, 2015.

      [9] Abriyono and A. Harjoko (2012), “Pengenalan Ucapan Suku Kata Bahasa Lisan Menggunakan Ciri LPC, MFCC, dan JST,” Indonesian Journal Computing and Cybernetics Systems, vol. 6, no. 2, pp. 23–34. DOI: 10.22146/ijccs.2149.

      [10] Nafisah S, Wahyunggoro O, and Nugroho LE (2016), “An Optimum Database for Isolated Word in Speech Recognition System,” Telkomnika, vol. 14, no. 2, pp. 588–597, 2016.

      [11] Fachrie M and Harjoko A, “using Elman Recurrent Neural Network Robust Indonesian Digit Speech Recognition using Elman Recurrent Neural Network,” in Prosiding Konferensi Nasional Informatika (KNIF), 2015, no. March, pp. 49–54.

      [12] Fugal DL (2009), Conceptual Wavelets in Digital Signal Processing. San Diego, California: Space & Signals Technical Publishing.

      [13] Biswas A, Sahu PK, Bhowmick A, and Chandra M (2015), “Admissible wavelet packet sub-band-based harmonic energy features for Hindi phoneme recognition,” IET Signal Processing Journal, vol. 9, no. 6, pp. 511–519.

      [14] Biswas A, Sahu PK, and Chandra M (2014), “Admissible wavelet packet features based on human inner ear frequency response for Hindi consonant recognition,” Computer Electronic Engineering, vol. 40, no. 4, pp. 1111–1122.

      [15] Chen YH and Yu SN (2006), “Comparison of DifferentWavelet Subband Features in the Classification of ECG Beats Using Probabilistic Neural Network,” in 2006 International Conference of the IEEE Engineering in Medicine and Biology Society, 2006, pp. 1398–1401.

      [16] Sunny S, Peter D, and Jacob KP (2013), “A Comparative Study of Wavelet Based Feature Extraction Techniques in Recognizing Isolated Spoken Words,” International Journal Signal Processing System, vol. 1, no. 1, pp. 49–53.

      [17] Kulkarni P, Kulkarni S, Mulange S, Dand A, and Cheeran AN (2014), “Support Vector Machines for Isolated Word Recognition using Wavelet Packet Features,” International Journal Engineering Technoogy. Res., no. 2, pp. 31–37.

      [18] Suyanto and Hartati S (2013), “Design of Indonesian LVCSR using Combined Phoneme The Approaches of LVCSR,” ICTS, pp. 191–196.

      [19] Gidas B and Murua A (1995), “Classification and clustering of stop consonants via nonparametric transformations and wavelets,” in 1995 International Conference on Acoustics, Speech, and Signal Processing, pp. 872–875.

      [20] Hidayat R, Priyatmadi, and Ikawijaya W (2015), “Wavelet based feature extraction for the vowel sound,” in 2015 International Conference on Information Technology Systems and Innovation (ICITSI), pp. 1–4. DOI: 10.1109/ICITSI.2015.7437702.

      [21] Kohavi R (1995), “A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection,”International Joint Conference on Artificial Intelligent, vol. 14, no. 12, pp. 1137–1143.

      [22] Chandra C and Yegnanarayana B (2002), “A constraint satisfaction model for recognition of stop consonant-vowel (SCV) utterances,” IEEE Transactions on Speech and Audio Processing, no. 7, pp. 472–480. DOI: 10.1109/TSA.2002.804298.




Article ID: 24076
DOI: 10.14419/ijet.v7i4.40.24076

Copyright © 2012-2015 Science Publishing Corporation Inc. All rights reserved.