Automatic extraction of eyes is a very important step in face detection and recognition system since eyes are one of the most stable features of the human face. In this paper, we present a novel technique using genetic programming for determining the classifier function to be used in the automatic detection of eyes in facial images. The feature terminals fed to the system are Gabor wavelet filtered image, mean, standard deviation and vertical position. Gabor wavelet transform has the optimal basis to extract local features. To find the Gabor wavelet to filter the image, we make use of Levenberg-Marquardt optimization. For the fitness function, we have used the concept of localization fitness, which is incorporated in the calculation of the precision and recall values to be included in fitness. We tested our system on the face images from the ORL databases and have presented our results. The result shows the effectiveness and flexibility provided by genetic programming in deciding the classifier for the detection of eyes in face images. © 2013 Springer International Publishing.
CITATION STYLE
Padole, C., & Athaide, J. (2013). Automatic eye detection in face images for unconstrained biometrics using genetic programming. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8298 LNCS, pp. 364–375). https://doi.org/10.1007/978-3-319-03756-1_33
Mendeley helps you to discover research relevant for your work.