Title: Study of geostatistical functions applied to automatic eye detection

Authors: J.D.S. Almeida; A.C. Silva; A.C. Paiva

Addresses: Applied Computing Group NCA/UFMA, Federal University of Maranhão, Av. dos Portugueses, SN, Campus do Bacanga, Bacanga 65085-580, São Luís, MA, Brazil. ' Applied Computing Group NCA/UFMA, Federal University of Maranhão, Av. dos Portugueses, SN, Campus do Bacanga, Bacanga 65085-580, São Luís, MA, Brazil. ' Applied Computing Group NCA/UFMA, Federal University of Maranhão, Av. dos Portugueses, SN, Campus do Bacanga, Bacanga 65085-580, São Luís, MA, Brazil

Abstract: Several computational systems that depend on the precise location of the eyes were developed in the last decades. Aware of this need, we propose a method for automatic eye detection in images of human faces using geostatistical functions (semivariogram, Getis index, Moran's index and Geary's coefficient) and support vector machines. The geostatistical measures are used as input features for a support vector machine classifier with the purpose of distinguishing patterns of eyes region and other areas of the face. The method was tested with ORL human face database, which contains 400 images of 40 persons, having ten different expressions for each person. The use of the proposed techniques showed to be very promising, since we obtained results of sensitivity of 92.2% for Moran's index, specificity of 93.4% and accuracy of 88.45% for semivariogram function.

Keywords: automatic eye detection; semivariogram; Getis index; Moran's index; geostatistical functions; innovative computing; ORL database; Geary's coefficient; support vector machine; SVM; pattern recognition; eyes; eye location.

DOI: 10.1504/IJICA.2012.050056

International Journal of Innovative Computing and Applications, 2012 Vol.4 No.3/4, pp.201 - 213

Received: 02 May 2011
Accepted: 22 Dec 2011
Published online: 22 Sep 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article