Estimação dos Parâmetros de uma SVM utilizando um Algoritmo Genético para o Reconhecimento de Caracteres Manuscritos

• Francisco Júnior UFC
• Kennedy Abreu UFC

Abstract

This work approaches an HCR system using a genetic algorithm to estimate the kernel parameters of an SVM for the recognition of the manuscript characters of the MNIST database with the objective of increasing the correct- ness percentage of the classifier by finding the best parameters. Digital image processing (PDI) techniques for image processing are also addressed. In the first topics are presented related introductory concepts and some works found in the literature. In the following topics, we describe the methodology used with the main concepts addressed, and the results achieved and conclusions about the approach taken.

References

Bonesso, D. (2013). Estimação dos Parâmetros do Kernel em um Classificador SVM na Classificação de Imagens Hiperespectrais em uma Abordagem Multiclasse. Master’s thesis, Universidade Federal do Rio Grande do Sul Centro Estadual de Sensoriamento Remoto e Meterologia Programa de Pós-Graduação em Sensoriameno Remoto, Porto Alegre.

Coelho, G. V. V. (2013). Seleção de Características usando Algoritmos Genéticos para Classificação de Imagens de Textos em Manuscritos e Impressos. Master’s thesis, Universidade Federal de Pernambuco, Recife.

Gonzalez, R. C. and Woods, R. E. (2006). Digital Image Processing (3rd Edition). Prentice-Hall, Inc., Upper Saddle River, NJ, USA.

Kumar, P., Sharma, N., and Rana, A. (2012). Handwritten character recognition using different kernel based svm classifier and mlp neural network (a comparison). International Journal of Computer Applications, 53(11).

LeCun, Y., Cortes, C., and Burges, C. J. (1998). The mnist database of handwritten digits. Disponível em: [https://goo.gl/LU1HuQ]. Acesso em: 20 de abril de 2017.

Linden, R. (2012). Algoritmos Genéticos. Editora Ciência Moderna, 3rd edition.

Mori, S., Suen, C. Y., and Yamamoto, K. (1995). Document image analysis. chapter Historical Review of OCR Research and Development, pages 244–273. IEEE Computer Society Press, Los Alamitos, CA, USA.

Pradeep, J., Srinivasan, E., and Himavathi, S. (2012). Neural network based recognition system integrating feature extraction and classification for english handwritten. International journal of Engineering, 25(2):99–106.

Rodrigues, R. J., Silva, E., and Thomé, A. C. G. (2001). Feature extraction using contour projection. Dispon´vel em: [https://goo.gl/M3YjoM]. Acesso em: 25 de julho de 2017.

Trier, Ø. D., Jain, A. K., and Taxt, T. (1996). Feature extraction methods for character recognition - a survey. Pattern Recognition (PR), 29(4):641–662.

Yerra, N., Varanasi, R., Adapaka, H., Surumalla, H., and Dantha, J. (2017). Recognition of handwritten characters using svm. International Journal of Innovate Research in Science and Engineering, 4(3).