Abstract
In this paper, we propose a novel HMM-based 2-D recognition engine, namely the NSHPZ-HMM. Like the reference model (the NSHP-HMM), the proposed classifier brings the efficient training and decoding algorithms of 1-D HMM to the 2-D modeling of spatial data. Furthermore, in contrast to the reference model which suffers from the short 2-D context limitation, our model uses the NSHP Markov random field to describe the contextual information at a ’zone’ level rather than a ’pixel’ level; the goal is to extend the context in order to give a better modeling of the spatial property of an image. Therefore, the use of high-level features extracted directly on the gray-level or color zones is possible, unlike what is done in a recognition based on classical NSHP-HMM, where the model, mandatorily, operates at a pixel level on normalized binary images; consequently, the applicability of our model is more general compared to the classical NSHP-HMM. Throughout this paper, we demonstrate the efficiency of the proposed approach at two stages. Firstly, in the theoretical study, we show the advantage of our model over other HMM-based 2-D classifiers; this part constitutes by itself, to our knowledge, the first complete overview of 2-D recognition approaches. Secondly, the experimental evaluation performed on recognition of handwritten digits/words provides the effectiveness of the NSHPZ-HMM against all other HMM-based 2-D recognizers and shows a good potential for other image recognition applications.
Similar content being viewed by others
Notes
We note that the ’site’ can corresponds to one pixel or a bloc of pixels.
Other studies such as 2[38] consider the order as the distance between sites. Thus, according to the latter view, the HMMRF defined here is a first-order Markovian model.
The corresponding solution for each drawback is noted by (’ ), i.e. (a’) is the proposed solution of drawback (a).
The site (i, j) corresponds to the pixel (i, j) for the NSHP-HMM model and to the zone \(Z_{ij}\) for the NSHPZ-HMM.
References
Agazzi OEx, Kuo S -S, Levin E, Pieraccini R (1993) Degraded text recognition using planar hidden markov models. In: 1993. ICASSP-93., 1993 IEEE International Conference on Acoustics, Speech, and Signal Processing. IEEE, vol 5, pp 113–116
Al-Ma’adeed S, Elliman D, Higgins C (2002) A data base for arabic handwritten text recognition research. In: 2002. Proceedings. Eighth International Workshop on Frontiers in Handwriting Recognition. IEEE, pp 485–489
Al-Nuzaili Q, Ali H, Siti H, Saeed F, Khalil MS (2017) An enhanced quadratic angular feature extraction model for arabic handwritten literal amount recognition. In: Recent Trends in Information and Communication Technology: Proceedings of the 2nd International Conference of Reliable Information and Communication Technology (IRICT 2017). Springer, vol 5, pp 369
Alma’adeed S, Higgins C, Elliman D (2004) Off-line recognition of handwritten arabic words using multiple hidden markov models. Knowl-Based Syst 17(2):75–79
Baggenstoss PM (2011) Two-dimensional hidden markov model for classification of continuous-valued noisy vector fields. IEEE Trans Aerosp Electron Syst 47(2):1073–1080
Boudaren MEY, Belaïd A (2009) Markov models and extensions for land cover mapping in aerial imagery. In: International Conference of Signal and Image Engineering-ICSIE, pp 2009
Boukerma H, Benouareth A, Farah N (2014) Nshp-hmm based on conditional zone observation probabilities for off-line handwriting recognition. In: 22nd International Conference on Pattern Recognition (ICPR). IEEE, pp 2961–2965
Boukerma H, Choisy C, Benouareth A, Farah N (2015) A performance evaluation of nshp-hmm based on conditional zone observation probabilities application to offline handwriting word recognition. In: 2015 13th International Conference on Document Analysis and Recognition (ICDAR). IEEE, pp 1091–1095
Boukerma H, Farah N (2010) A novel arabic baseline estimation algorithm based on sub-words treatment. In Frontiers in Handwriting Recognition (ICFHR) International Conference on. IEEE, pp 335–338
Cecotti H, Vajda S, Belaïd A (2005) High performance classifiers combination for handwritten digit recognition. In: Pattern Recognition and Data Mining. Springer, pp 619–626
Chevalier S, Geoffrois E, Prêteux F (2003) A 2d dynamic programming approach for markov random field-based handwritten character recognition. In: IAPR International Conference on Image and Signal Processing. Citeseer
Choisy C Dynamic handwritten keyword spotting based on the nshp-hmm. In: 2007. ICDAR 2007 Ninth International Conference on Document Analysis and Recognition. IEEE, vol 1, pp 242–246
Choisy C, Belaïd A (2002) Cross-learning in analytic word recognition without segmentation. Int J Doc Anal Recognit 4(4):281–289
Choisy C et al (2003) Coupling of a local vision by markov field and a global vision by neural network for the recognition of handwritten words. In: null. IEEE, pp 849
Devijver PA, Dekesel M (1988) Champs aléatoires de pickard et modélisation d’images digitales. Traitement Signal 5(5):131–150
Devijver PA, Dekesel MM (1987) Learning the parameters of a hidden markov random field image model: A simple example. In: Pattern Recognition Theory and Applications. Springer, pp 141–163
El-Melegy MT, Abdelbaset AA (2007) Global features for offline recognition of handwritten arabic literal amounts. In: 2007. ICICT ITI 5th International Conference on Information and Communications Technology. IEEE, pp 125–129
Feng Q, Minghua D, Minping Q, Xueqing Z (2000) A novel algorithm for handwritten chinese character recognition. In: Advances in Multimodal Interfaces — ICMI. Springer, pp 379–385
Fink GA (2008) Markov models for pattern recognition: from theory to applications. Springer Science & Business Media, Berlin
Graves A (2008) Supervised Sequence Labelling with Recurrent Neural Networks. PhD thesis, Technische Universitat Munchen, Fakultat fur Informatik, pp 7
Graves A, Fernández S, Schmidhuber J (2007) Multi-dimensional recurrent neural networks. In: Artificial Neural Networks – ICANN, pp 549–558
Graves A, Schmidhuber J (2009) Offline handwriting recognition with multidimensional recurrent neural networks. In: Advances in neural information processing systems, pp 545–552
Impedovo D, Pirlo G (2014) Zoning methods for handwritten character recognition A survey. Pattern Recogn 47(3):969–981
Jeng F-C, Woods JW (1987) On the relationship of the markov mesh to the nshp markov chain. Pattern Recogn Lett 5(4):273–279
Joshi D, Li J, Wang JZ (2006) A computationally efficient approach to the estimation of two-and three-dimensional hidden markov models. IEEE Trans Image Process 15(7):1871–1886
Kurata D, Nankaku Y, Tokuda K, Kitamura T, Ghahramani Z (2006) Face recognition based on separable lattice hmms. In: 2006. ICASSP 2006 Proceedings IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, vol 5, pp V–V
Lazzerini B, Marcelloni F (2001) A fuzzy approach to 2d-shape recognition. IEEE Trans Fuzzy Syst 9 (1):5–16
LeCun Y, Bottou L, Bengio Y, Haffner P (1998) Gradient-based learning applied to document recognition. Proc IEEE 86(11):2278–2324
Li J, Najmi A, Gray RM (2000) Image classification by a two-dimensional hidden markov model. IEEE Trans Signal Process 48(2):517–533
Likforman-Sulem L, Sigelle M (2007) Recognition of degraded handwritten digits using dynamic bayesian networks. In: Document Recognition and Retrieval XIV, DRR 2007
Likforman-Sulem L, Sigelle M (2008) Recognition of degraded characters using dynamic bayesian networks. Pattern Recogn 41(10):3092–3103
Ma X, Schonfeld D, Khokhar A (2007) A general two-dimensional hidden markov model and its application in image classification. In: 2007. ICIP 2007. IEEE International Conference on Image Processing. IEEE, vol 6, pp VI-41
Ma X, Schonfeld D, Khokhar A (2008) Image segmentation and classification based on a 2d distributed hidden markov model. In Visual Communications and Image Processing 2008. International Society for Optics and Photonics, vol 6822, pp 68221F
Madhogaria S, Baggenstoss PM, Schikora M, Koch W, Cremers D (2015) Car detection by fusion of hog and causal mrf. IEEE Trans Aerosp Electron Syst 51(1):575–590
Othman H, Aboulnasr T (2003) A separable low complexity 2d hmm with application to face recognition. IEEE Trans Pattern Anal Mach Intell 25(10):1229–1238
Park H-S, Lee S-W (1998) A truly 2-d hidden markov model for off-line handwritten character recognition. Pattern Recogn 31(12):1849–1864
Park H-S, Sin B-K, Moon J, Lee S-W (2001) A 2-d hmm method for offline handwritten character recognition. Int J Pattern Recogn Artif Intell 15(01):91–105
Perronnin F (2004) A probabilistic model of face mapping applied to person recognition. PhD thesis, ECOLE POLYTECHNIQUE
Perronnin F, Dugelay J-L, Rose K (2003) Iterative decoding of two-dimensional hidden markov models. In: 2003. proceedings.(ICASSP’03) IEEE International Conference on Acoustics, Speech, and Signal Processing. IEEE, vol 3, pp III–329
Rabiner LR (1989) A tutorial on hidden markov models and selected applications in speech recognition. Proc IEEE 77(2):257–286
Ronee MA, Uchida S, Sakoe H (2001) Handwritten character recognition using piecewise linear two-dimensional warping. In: 2001. Proceedings. Sixth International Conference on Document Analysis and Recognition. IEEE, pp 39-43
Saon G (1999) Cursive word recognition using a random field based hidden markov model. Int J Doc Anal Recognit 1(4):199–208
Saon G, Belaïd A (1997) High performance unconstrained word recognition system combining hmms and markov random fields. Int J Pattern Recognit Artif Intell 11(05):771–788
Shenoy R, Shih M-C, Rose K (2016) Deformable registration of biomedical images using 2d hidden markov models. IEEE Trans Image Process 25(10):4631–4640
Tamamori A, Nankaku Y, Tokuda K (2014) Image recognition based on separable lattice trajectory 2-d hmms. IEICE Trans Inf Syst 97(7):1842–1854
Uchida S, Sakoe H (1999) An efficient two-dimensional warping algorithm. IEICE Trans Inf Syst 82 (3):693–700
Uchida S, Sakoe H (2005) A survey of elastic matching techniques for handwritten character recognition. IEICE Trans Inf Syst 88(8):1781–1790
Vajda S, Belaïd A (2005) Structural information implant in a context based segmentation-free hmm handwritten word recognition system for latin and bangla script. In: 2005. Proceedings. Eighth International Conference on Document Analysis and Recognition. IEEE, pp 1126–1130
Wan W, Yuan L, Zhao Q, Fang T (2018) Two-dimensional hidden semantic information model for target saliency detection and eyetracking identification. J Electron Imaging 27(1):013006
Wang G-G, Gan Z-L, Tang G-j, Cui Z-g, Zhu X-c (2016) Basic problems solving for two-dimensional discrete 3× 4 order hidden markov model. Chaos Solitons Fractals 89:73–82
Wang G-g, Tang G-j, Gan Z-l, Cui Z-g, Zhu X-c (2016) Basic problems and solution methods for two-dimensional continuous 3× 3 order hidden markov model. Chaos Solitons Fractals 89:435–446
Wang Q, Zhao R, Chi Z, Feng DD (2000) Hmmrf: a stochastic model for offline handwritten chinese character recognition. In: 2000. WCCC-ICSP 2000. 5th International Conference on Signal Processing Proceedings. IEEE, vol 3, pp 1475–1478
Yujian L (2007) An analytic solution for estimating two-dimensional hidden markov models. Appl Math Comput 185(2):810–822
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Boukerma, H., Choisy, C., Farah, N. et al. The efficiency of the NSHPZ-HMM: theoretical and practical study. Appl Intell 48, 4660–4677 (2018). https://doi.org/10.1007/s10489-018-1217-z
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-018-1217-z