Module DetectorDOG
source code
Purpose: Find regions of interest using the multiscale differance of
gaussians opereator.
Inspired by: M. Brown and D. Lowe. Invariant features from interest
point groups. British Machine Vision Conference. 2002.
Contributed by: David Bolme 2008.
|
|
TaylorFit(block)
a*s*s + b*x*x + c*y*y + d*x*y + e*s + f*x + g*y + h |
source code
|
|
|
|
TaylorSubpixel(params)
a*s*s + b*x*x + c*y*y + d*x*y + e*s + f*x + g*y + h |
source code
|
|
|
|
TaylorApprox(params,
s,
x,
y)
a*s*s + b*x*x + c*y*y + d*x*y + e*s + f*x + g*y + h |
source code
|
|
|
|
EXTREMA_SIZE = 3
|
|
|
DEFAULT_SIGMA = 1.4142135623730951
|
|
|
DEFAULT_TYPE = 'd' |
|
|
DOG_SCALES = 2
|
|
|
MAX_EXTREMA = 1024
|
|
|
__package__ = 'pyvision.point' |