Revista ELECTRO
Vol. 41 – Año 2019
Artículo
TÍTULO
Red Neuronal Convolucional de Bajo Costo para el Reconocimiento de Iris
AUTORES
González Estrada Adrián Alberto, Ramírez Quintana Juan Alberto y Vega Pineda Javier
RESUMEN
Se presenta la arquitectura de una red neuronal convolucional (CNN) de bajo costo basad a en el modelo AlexNet y se comparó su rendimiento con el de las arquitecturas CNN AlexNet (sin modificar) y la CNN VGG-16. El reconocimiento de iris se refiere al proceso automático de reconocer individ uos en base a sus patrones visuales específicos de su iris. La naturaleza de las características aleatorias y distintivas del iris lo convierte en una señal adecuada para el reconocimiento biométrico. Avances recientes en el área de aprendizaje profundo ("deep learning") indican que las CNN extraen características que representan y generan información importante existente en las imá genes. Por lo cual, s e trabajó en el desarrollo de un modelo basado en la CNN AlexNet y el modelo propuesto se pr obó con 15 clases seleccionadas de la base de datos U TIRIS V1 (imágenes de ojos humanos) para evaluar su desempeñ o y complejidad computacional. Para este propósito, se introdu cen imágenes segmentadas y normalizadas sin ruido, logrando buenos resultados comparativos con respecto a las antes mencionadas CNN.
Palabras Clave: Palabras C lave: Reconocimiento de iris, red neuronal convolucional (CNN), aprendizaje profundo.
ABSTRACT
An architecture of a low-cost convolutional neural network (CNN) model based on the AlexNet model is presen ted and its performance was compared with the CNN AlexNet (unmodified) and the CNN VGG-16 architectures. Iris recognition refers to the automatic process of recognizing individuals based on their iris specific visual patterns. The nature of the random and distinctive features of the iris makes it an adequate signal for biometric recognition. Recent advances i n the deep learning area indicate that CNNs extract characteristics that can represent and generate important information existing in images. Therefore, we worked on the development of a model based on the CNN AlexNet and the proposed model was tested with 15 selected classes of the UTIRIS V1 database (images of human eyes) to evaluate its performance and computational complexity. For that purpose, we introduce segmented and normalized images without noise, achieving good comparative results with respect to the CNN.
Keywords: Iris recognition, convolutional neural networks (CNN ), deep learning.1
REFERENCIAS
[1] M. Oravec, “Feature extraction and classification by machine learning methods for biometric recog nition of face and iris,” ELMAR 56th Int. Symp., September, pp. 1 –4, 2014.
[2] J. Daugman, “How Iris Recognition Works,” Essent. Guid. to Image Process., vol. 41, no. 1, pp. 715 –739, 2009.
[3] M. Shamsi, P. Saad, and A. Rasouli, “Iris segmentation and normalization approach,” J. Teknol. Mklm., vol. 19, no. 2, p p. 88 –101, 2008.
[4] M. G Alaslani and L. A. Elrefaei, “Convolutional Neural Network Based Feature Extraction for IRIS Recognition,” Int. J. Comput. Sci. Inf. Technol., vol. 10, no. 2, pp. 65 –78, 2018.
[5] A. Gangwar and A. Joshi, “Deepirisnet: Deep Iris Representation With Applications In Iris Recognition And Cross-Sensor Iris Recognition,” 2016 IEEE Int. Conf. Image Process, vol. 2015-Novem ber 2016.
[6] F.F. Khan et al., "Iris Recognition using Machine Learnin g from Smartphone Captured Images in Visible Light," ELMAR, 56th Int. Symp., vol. 7, pp. 26-28, September 2017.
[7] Z. Othman and A. Satria Prabuwono, “Preliminary study on iris recognition system: Tissues of body organs in iridology,” Proc. IEEE EMBS Conf. Bi omed. Eng. Sci. IECBES, pp. 115 –119, December 2010.
[8] S. El-Naggar and A. Ross, “Which dataset is this iris image from?,” IEEE Int. Work. Inf. Forensics Secur. WIFS 2015-Proc., November, 2015.
[9] M.S. Hosseini, B.N. Arabbi, and H. Soltanian-Zadeh, "Pigment me lanin: Pattern for iris recgnition," IEEE Trans. Instrum. Meas., vol. 59, no. 4, pp. 792-804, 2010.
[10] F. F. Khan, A. Akif, and M. A. Haque, “Iris Recognition using Machine Learning from Smartphone Captured Images in Visible Light,” IEEE Intl. Conf. Telecomm. and Photonics (ICTP), pp. 26 –28, 2017.
[11] A. S. Al-Waisy, R. Qahwaji, S. Ipson, S. Al-Fahdawi, and T. A. M. Nagem, “A multi-biometric iris recognition system based on a deep learning approach,” Pattern Anal. Appl., vol. 21, no. 3, pp. 783 –802, 2018.
[12] G. E. H. Alex Krizhevsky, Ilya Sutskever, “ImageNet Classification with Deep Convolutional Neural Networks,” LSVRC-2010, vol. 12, 2012.
[13] O. Oyedotun and A. Khashman, “Iris nevus diagnosis: Convolutional neural network and deep belief network,” Turkish J. Electr. Eng. Comput. Sci., vol. 25, no. 2, pp. 1106 –1115, 2017.
[14] S. Ioffe and C. Szegedy, “Batch normalization: Accelerating deep network t raining by reducing internal covariate shift,” arXiv Prepr. arXiv1502.03167, 2015.
[15] K. Wang and A. Kumar, “Cross-spectral iris recognition using CNN and supervised discrete hashing,” Pattern Recognit., vol. 86, pp. 85 –98, 2019.
[16] W. Zhang, C. Wang, and P. Xue, “Application of convolution neural network in Iris recognition technology,” 4th Int. Conf. Syst. Informatics, ICSAI 2017, pp. 1169 –1174, January 2018.
[17] F. Ma, G. Po, S. Sa, and L. Ve, “A deep learning approach for iris sensor model identificatio n,” vol. 4 6, no. 3, pp. 1 –8, 2016.
CITAR COMO:
González Estrada Adrián Alberto, Ramírez Quintana Juan Alberto y Vega Pineda Javier, "Red Neuronal Convolucional de Bajo Costo para el Reconocimiento de Iris", Revista ELECTRO, Vol. 41, 2019, pp. 142-147.
VERSIÓN PDF