Combinatorial Optimization Approach for Arabic Word Recognition Based on Adaptive Simulated Annealing

  • Zeineb ZouaouiEmail author
  • Imen Ben CheikhEmail author
  • Mohamed JemniEmail author
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11896)


The present paper proposes an approach based on a combinatorial optimization technique for Arabic word recognition, distinguished by its flexional nature and significant topological variability. We treat a large vocabulary of Arabic decomposable words, which we choose to factorize them by their roots and schemes. We adopt a structure that resembles a molecular cloud. This design rhymes well with the Arabic linguistic philosophy of constructing words from their roots. Each sub-vocabulary, corresponding to a sub-cloud, embodies neighboring words, which are derived from one root and follow different schemes and forms of derivation, flexion, and agglutination (proclitic and enclitic). Therefore, we propose to use the metaheuristic simulated annealing (SA) method, as a recognition approach, in this wide cloud. It’s an algorithm based on elastic comparisons between their structures and primitives. As an extension of previous works, we opt to implement the SA algorithm by integrating linguistic knowledge. Preliminary experiments were conducted on Arabic word corpus including samples and agglutinated words from APTI database and yielded interesting outcomes.


Simulated annealing Levenshtein distance Morphological peculiarities Combinatorial optimization APTI database 


  1. 1.
    Touj, S., Ben Amara, N., Amiri, H.: A hybrid approach for off-line Arabic handwriting recognition based on a planar Hidden Markov Modeling. In: ICDAR 2007, Brazil, pp. 964–968 (2007)Google Scholar
  2. 2.
    Avila, J.M.: Optimisation de modèles markoviens pour la reconnaissance de l’écrit. Ph.D., University of Rouen (1996)Google Scholar
  3. 3.
    Cheriet, M., Beldjehem, M.: Visual processing of Arabic handwriting: challenges and new directions. In: SACH 2006, India, pp. 1–21 (2006)Google Scholar
  4. 4.
    Kanoun, S., Alimi, A.M., Lecourtier, Y.: Natural language morphology integration in off-line Arabic optical text recognition. IEEE Trans. Syst. Man Cybern.—Part B: Cybern. 41(2), 579–590 (2011)CrossRefGoogle Scholar
  5. 5.
    Ben Cheikh, I., Allagui, I.: Planar Markovian approach for the recognition of a wide vocabulary of Arabic decomposable words. In: ICDAR, pp. 1031–1035 (2015)Google Scholar
  6. 6.
    Levenshtein, V.: Binary codes capable of correcting deletions, insertions and reversals. SOL Phys. Dokl. 10, 707–710 (1966)MathSciNetGoogle Scholar
  7. 7.
    Anigbogu, J.: Reconnaissance de textes imprimés multifontes àl’aide de modèles stochastiques et métriques. Ph.D., University of Nancy 1 (1992)Google Scholar
  8. 8.
    Rani, S., Singh, J.: Enhancing Levenshtein’s edit distance algorithm for evaluating document similarity. In: Sharma, R., Mantri, A., Dua, S. (eds.) ICAN 2017. CCIS, vol. 805, pp. 72–80. Springer, Singapore (2018). Scholar
  9. 9.
    Huang, K., Hsieh, Y.: Very fast simulated annealing for pattern detection and seismic. In: IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2011, Vancouver, BC, Canada, 24–29 July 2011Google Scholar
  10. 10.
    Goyal, A., Sourav, P.A., Thangavelu, A.: A comparative analysis of simulated annealing based intuitionistic fuzzy k-mode algorithm for clustering categorical data. Int. J. Comput. Inf. Syst. Ind. Manag. Appl. 9, 232–240 (2017)Google Scholar
  11. 11.
    Baptiste, A.: Les métaheuristiques en optimisation combinatoire. Thesis to obtain probative exam on computing, National Conservatory of Arts and Crafts, Paris (2006)Google Scholar
  12. 12.
    Gueddah, H.: La correction orthographique des textes arabes: Contribution àla résolution d’ordonnancement et de l’insuffisance des lexiques. Ph.D. University of Mohamed V, Rabat (2017)Google Scholar
  13. 13.
    Carbonnel, S., Anquetil, E.: Apprentissage automatique d’une distance d’édition dédié à la reconnaissance d’écriture manuscrite. In: CIFED (2004)Google Scholar
  14. 14.
    Xinchao, Z.: Simulated annealing algorithm with adaptive neighborhood. Appl. Soft Comput. 11, 1827–1836 (2011)CrossRefGoogle Scholar
  15. 15.
    Ben Cheikh, I., Zouaoui, Z.: HMM based classifier for the recognition of roots of a large canonical Arabic vocabulary. In: ICPRAM, pp. 244–252 (2013)Google Scholar
  16. 16.

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Latice LaboratoryENSIT, University of TunisTunisTunisia

Personalised recommendations