-
Notifications
You must be signed in to change notification settings - Fork 3
/
steerableDetector.m
27 lines (25 loc) · 1.44 KB
/
steerableDetector.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
%[res, theta, nms, rotations] = steerableDetector(img, M, sigma) performs edge/ridge detection through a generalization of Canny's alorithm based on steerable filters
%
% Inputs:
% img : input image
% M : order of the filter, between 1 and 5
% : Odd orders: edge detectors, M = 1 is equivalent to Canny's detector
% : Even orders: ridge detectors
% Higher orders provide better orientation selectivity and are less sensitive to noise,
% at a small trade-off in computational cost.
% sigma : standard deviation of the Gaussian kernel on which the filters are based
% {nAngles} : optional input specifying the number of angles computed for the filterbank output. Default: 36
%{bordercond} : optional character input specifying the border condition
% : 'mirror' (default), 'replicate', 'periodic', 'zeros'
%
% Outputs:
% res : response to the filter
% theta : orientation map
% nms : non-maximum-suppressed response
% rotations : response of the input to rotated versions of the filter, at 'nAngles' different angles.
%
% For more information, see:
% M. Jacob et al., IEEE Trans. Image Proc., 26(8) 1007-1019, 2004.
% Copyright (C) 2011-2012 Francois Aguet
% Adapted from the SteerableJ package, Copyright (C) 2005-2008 Francois Aguet, Biomedical Imaging Group, EPFL.
function [res, theta, nms, filterBank] = steerableDetector(img, M, sigma) %#ok<STOUT,INUSD>