how to detect eye location?
Ältere Kommentare anzeigen
Antworten (1)
You could follow the following general approach. Note that a lot will depend on your image, noise and kind of data.
Use imfindcirciles to find circles in the correct radius range. If there are multiple matches, adjust parameters and/or use additional post-processing like some form of thresholding and morphology. The centers returned by imfindcircles are your eye centers.
Read the following for an understanding of imfindcircles and its parameters.
9 Kommentare
KHAN
am 23 Mär. 2013
Image Analyst
am 23 Mär. 2013
imfindcircle() can deal with circles that are clipped or flat on one side, or are partial circles (dumbbell shapes).
KHAN
am 25 Mär. 2013
Image Analyst
am 25 Mär. 2013
Contact the Mathworks and upgrade to the latest version. Why use a version that's three years out of date, and that's missing a critical function that could help you greatly?
KHAN
am 27 Mär. 2013
Image Analyst
am 27 Mär. 2013
KHAN
am 28 Mär. 2013
Image Analyst
am 28 Mär. 2013
I'm sure there must be lots of them listed here: http://iris.usc.edu/Vision-Notes/bibliography/contents.html
Kategorien
Mehr zu Hough Transform finden Sie in Hilfe-Center und File Exchange
Produkte
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!