sift_match.m

function [X, Y, S1, S2] = sift_match(I1, I2, theta)
%   [X, Y, S1, S2] = SIFT_MATCH(I1, I2, THETA) does sift match use
%   VL_Feat. 
%
% Input:
%   I1, I2: Tow input image.
%
%   theta: A descriptor D1 is matched to a descriptor D2 only if the
%       distance d(D1,D2) multiplied by THRESH is not greater than the
%       distance of D1 to all other descriptors. The default value of
%       THRESH used by VL_Feat is 1.5.
%
% Output:
%   X, Y: SIFT matches of interest points.
%   S1, S2: Scale of the matches point pairs

if size(I1, 3)>1
    [f1,d1] = vl_sift(im2single(rgb2gray(I1))) ;
else
    [f1,d1] = vl_sift(im2single(I1)) ;
end
if size(I2, 3)>1
    [f2,d2] = vl_sift(im2single(rgb2gray(I2))) ;
else
    [f2,d2] = vl_sift(im2single(I2)) ;
end

[matches, scores] = vl_ubcmatch(d1,d2, theta) ;

x1 = f1(1,matches(1,:)) ;
x2 = f2(1,matches(2,:)) + size(I1,2) ;
y1 = f1(2,matches(1,:)) ;
y2 = f2(2,matches(2,:)) ;
X = [x1; y1]';
Y = [f2(1,matches(2,:)); y2]';
S1 = f1(3, matches(1,:));
S2 = f2(3, matches(2,:));

% figure ; clf ;
% imagesc(cat(2, I1, I2)) ;hold on ;
% h = line([x1 ; x2], [y1 ; y2]) ; set(h,'linewidth', 2, 'color', 'b') ;
% axis equal ;axis off  ;