This paper presents a method for eyes location using a two-level hierarchy of SVM (Support Vector Machines) classifiers. On the first level, a two-eye region classifier is obtained by training the SVM using grayscale projections of the two-eye region images. Utilizing this classifier, the region where the two eyes lie can be located by searching the whole face image. On the second level, the left and right eye classifier are obtained by training SVM using grayscale of left and right eye images respectively. Using these two classifiers, the two eyes can be precisely located by searching the output region of the first level. Experimental results show that this method is sufficiently generic and can cope with more various image conditions than exiting techniques. © Springer-Verlag 2004.
CITATION STYLE
Li, Y., & Ou, Z. (2004). Eyes location by hierarchical SVM classifiers. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3173, 611–615. https://doi.org/10.1007/978-3-540-28647-9_100
Mendeley helps you to discover research relevant for your work.