Abstract
In this paper we present a robust polynomial classifier based on L 1-norm minimization. We do so by reformulating the classifier training process as a linear programming problem. Due to the inherent insensitivity of the L 1-norm to influential observations, class models obtained via L 1-norm minimization are much more robust than their counterparts obtained by the classical least squares minimization (L 2-norm). For validation purposes, we apply this method to two recognition problems: character recognition and sign language recognition. Both are examined under different signal to noise ratio (SNR) values of the test data. Results show that L 1-norm minimization provides superior recognition rates over L 2-norm minimization when the training data contains influential observations especially if the test dataset is noisy.
Similar content being viewed by others
References
Sanderson C, Bengio S (2003) Robust features for frontal face authentication in difficult image conditions. In: Proc IDIAP-RR 03-05, Martigny, Switzerland
Comaniciu D, Meer P (1997) Robust, analysis of feature spaces: color image segmentation. In: Proc IEEE conf on comp vis and pattern recognition, Puerto Rico, pp 750–755
Zheng Y, Li H, Doermann D (2004) Machine printed text and handwriting identification in noisy document images. IEEE Trans Pattern Anal Mach Intell 26:337
Addison WD, Glendinning RH (2006) Robust image classification. Signal Process 86(7):1488–1501
Zhu Q, Alwan A (2003) Non-linear feature extraction for robust speech recognition in stationary and non-stationary noise. Comput Speech Lang 17(4):381–402
Assaleh K, Mammone R (1994) LP-derived features for speaker identification. IEEE Trans Speech Audio Process 2(4):630–638
Ming T, Hazen J, Glass J, Reynolds D (2007) Robust speaker recognition in noisy conditions. IEEE Trans Speech Audio Process 15(5):1711–1723
Gales M, Young S (1996) Robust continuous speech recognition using parallel model combination. IEEE Trans Speech Audio Process 4(5):352–359
Drygajlo A, Virag N, Cosendai G (1995) Robust speech recognition in noise using speech enhancement based on masking properties of the auditory system and adaptive HMM. In: Proc of the 4th European conference on speech communication and technology, Madrid, Spain, pp 473–476
Campbell WM, Assaleh KT, Broun CC (2002) Speaker recognition with polynomial classifiers. IEEE Trans Speech Audio Process 10(4):205–212
Assaleh KT, Campbell WM (1999) Speaker identification using a polynomial-based classifier. In: Proc of the fourth international symposium on signal processing and its applications ISSPA’99. Brisbane, Australia, August 1999
Campbell WM, Assaleh KT (1999) Low-complexity small-vocabulary speech recognition for portable devices. In: Proc of the fourth international symposium on signal processing and its applications ISSPA’99. Brisbane, Australia, August 1999
Assaleh K, Al-Nashash H (2005) A novel technique for the extraction of fetal ECG using polynomial networks. IEEE Trans Biomed Eng 52(6):1148–1152
Fukunaga K (1990) Introduction to statistical pattern recognition. Academic Press, San Diego
Schurmann J (1996) Pattern classification. Wiley, New York
Golub G, Van Loan C (1989) Matrix computations. John Hopkins Press, Baltimore
Cadzow J (2002) Minimum ℓ1,ℓ2, and ℓ∞ norm approximate solutions to an overdetermined system of linear equations. Digit Signal Process 12(4):524–560
Vanderbei R (2001) Linear programming foundations and extensions, 2nd edn. Kluwer Academic, Dordrecht
Dax A (2006) The ℓ1 solution of linear inequalities. Comput Stat Data Anal 50(1):40–60
Shanableh T, Assaleh K, Al-Rousan M (2007) Spatio-temporal feature extraction techniques for isolated Arabic sign language recognition. IEEE Trans Syst Man Cybern, Part B Cybern 37(3) 641–650
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Assaleh, K., Shanableh, T. Robust polynomial classifier using L 1-norm minimization. Appl Intell 33, 330–339 (2010). https://doi.org/10.1007/s10489-009-0169-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-009-0169-8