A Visual Model based Edge Detection Algortithm

SEMINARIO DE IMÁGENES Y VISIÓN
INSTITUTO DE OPTICA (CSIC)
A Visual Model based Edge Detection Algortithm
Dr. Eli Peli
Schepens Eye Research Institute, Harvard University
ABSTRACT
An algorithm for the detection of visually relevant luminance features is presented. The algorithm is
motivated and directed by current models of the visual system. The image is first band-pass filtered
with oriented filters at a number of scales an octave apart. The supra threshold image contrast details
at each scale are then identified and are compared across scales to find locations in which the signal
polarity (sign) is identical at all scales, representing a minimal level of phase congruence across scales.
The result is a polarity preserving features map representing the edges with pairs of light and dark
curves on corresponding sides of the contour. The algorithm is implemented without free (fitted)
parameters. All parameters are directly derived from visual models and from measurements on human
observers. The algorithm is shown to be robust with respect to variations in the filters' parameters and
requires no use of quadrature filters or Hilbert transforms.
Miércoles, 27 junio 2001
12:30 horas
Sala de Conferencias. Instituto de Optica (CSIC) .
C/ Serrano 121, 28006 Madrid
Información:
Susana Marcos
Instituto de Optica, CSIC
Tel: 915616800 x2306; email: [email protected]