Presentation 1996/1/26
A Neural Method for Localizing Eyes in Facial Images
Kazunori Kato, Koichiro Kawakami, Takashi Hamada,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a method for localizing eyes in a human facial image, using a neural model for simple cells in the visual cortex. The method consists of two stages. First, candidates for the eye positions are extracted by a network of Gabor filters: receiving excitatory inputs from the filters with central frequencies and orientations matched with the eye regions and inhibitory inputs from those with lower frequencies, the network yields output with threshold. Second, exact eye positions are determined by minimizing a cost function on positional relationship among the eyes and the eyeblows. We validate the method for facial images of 75 persons, within a certain variation of head orientation and size.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) simple cells / localization of the eyes / Gabor function
Paper # HIP95-31
Date of Issue

Conference Information
Committee HIP
Conference Date 1996/1/26(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Vice Chair

Paper Information
Registration To Human Information Processing (HIP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Neural Method for Localizing Eyes in Facial Images
Sub Title (in English)
Keyword(1) simple cells
Keyword(2) localization of the eyes
Keyword(3) Gabor function
1st Author's Name Kazunori Kato
1st Author's Affiliation Kobe University of Mercantile Marine()
2nd Author's Name Koichiro Kawakami
2nd Author's Affiliation Faculty of Engineering, Kobe University
3rd Author's Name Takashi Hamada
3rd Author's Affiliation Electrotechnical Laboratory
Date 1996/1/26
Paper # HIP95-31
Volume (vol) vol.95
Number (no) 499
Page pp.pp.-
#Pages 6
Date of Issue